Attachment 123136 Details for Bug 94955 – gallivm_debug shaders (inline sampler)

gallivm_debug shaders (inline sampler)

dual-depth-peel.txt (text/plain), 354.33 KB, created by Bruce Cherniak on 2016-04-22 00:07:34 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Bruce Cherniak

Created: 2016-04-22 00:07:34 UTC

Size: 354.33 KB

patch

obsolete

>UpdateCTestConfiguration  from :/home/bruce/Projects/VTK/build/DartConfiguration.tcl
>Parse Config file:/home/bruce/Projects/VTK/build/DartConfiguration.tcl
> Add coverage exclude regular expressions.
> Add coverage exclude: vtk.*TCLInit.cxx
> Add coverage exclude: vtk[^\.]+(Java|Python|Tcl).cxx
> Add coverage exclude: .*vtkOpenGLState.*
> Add coverage exclude: .*Testing.Cxx.*cxx
> Add coverage exclude: .*Testing.Cxx.*h
> Add coverage exclude: .*moc_.*cxx
> Add coverage exclude: .*/Rendering/OpenGL/vtkgl.*
> Add coverage exclude: .*/Utilities/.*
> Add coverage exclude: .*/ThirdParty/.*
> Add coverage exclude: .*vtkOpenGLPolyDataMapper.*
>UpdateCTestConfiguration  from :/home/bruce/Projects/VTK/build/DartConfiguration.tcl
>Parse Config file:/home/bruce/Projects/VTK/build/DartConfiguration.tcl
>Test project /home/bruce/Projects/VTK/build
>Constructing a list of tests
>Done constructing a list of tests
>Checking test dependency graph...
>Checking test dependency graph end
>test 1525
>    Start vtkRenderingCoreCxx-TestTextureRGBADepthPeeling
>
>Test command: /home/bruce/Projects/VTK/build/bin/vtkRenderingCoreCxxTests "TestTextureRGBADepthPeeling" "-D" "/home/bruce/Projects/VTK/build/ExternalData//Testing" "-T" "/home/bruce/Projects/VTK/build/Testing/Temporary" "-V" "/home/bruce/Projects/VTK/build/ExternalData/Rendering/Core/Testing/Data/Baseline/TestTextureRGBADepthPeeling.png"
>Test timeout computed to be: 3600
>SWR detected AVX2
>asked to convert unsupported format PIPE_FORMAT_X8Z24_UNORM
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>Mesa warning: couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
>asked to convert unsupported format PIPE_FORMAT_A8B8G8R8_SRGB
>asked to convert unsupported format PIPE_FORMAT_NONE
>Unexpected PIPE_CAP 125 query
>
>; Function Attrs: alignstack(32)
>define void @VS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx) #0 {
>entry:
>  %temp10 = alloca <8 x float>
>  %temp9 = alloca <8 x float>
>  %temp8 = alloca <8 x float>
>  %temp = alloca <8 x float>
>  %output7 = alloca <8 x float>
>  %output6 = alloca <8 x float>
>  %output5 = alloca <8 x float>
>  %output4 = alloca <8 x float>
>  %output3 = alloca <8 x float>
>  %output2 = alloca <8 x float>
>  %output1 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 0
>  %num_vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 1
>  %0 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 0
>  %1 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %0
>  %2 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 0
>  %3 = load <8 x float>, <8 x float>* %2
>  %4 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 1
>  %5 = load <8 x float>, <8 x float>* %4
>  %6 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 2
>  %7 = load <8 x float>, <8 x float>* %6
>  %8 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 3
>  %9 = load <8 x float>, <8 x float>* %8
>  %10 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 0
>  %11 = load <8 x float>, <8 x float>* %10
>  %12 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 1
>  %13 = load <8 x float>, <8 x float>* %12
>  %14 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 2
>  %15 = load <8 x float>, <8 x float>* %14
>  %16 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 3
>  %17 = load <8 x float>, <8 x float>* %16
>  %18 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 2
>  %19 = load i32, i32* %18
>  %20 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 3
>  %21 = load <8 x i32>, <8 x i32>* %20
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output1
>  store <8 x float> zeroinitializer, <8 x float>* %output2
>  store <8 x float> zeroinitializer, <8 x float>* %output3
>  store <8 x float> zeroinitializer, <8 x float>* %output4
>  store <8 x float> zeroinitializer, <8 x float>* %output5
>  store <8 x float> zeroinitializer, <8 x float>* %output6
>  store <8 x float> zeroinitializer, <8 x float>* %output7
>  %"&vs_constants[]" = getelementptr [32 x float*], [32 x float*]* %vs_constants, i32 0, i32 0
>  %"vs_constants[]" = load float*, float** %"&vs_constants[]"
>  %"&num_vs_constants[]" = getelementptr [32 x i32], [32 x i32]* %num_vs_constants, i32 0, i32 0
>  %"num_vs_constants[]" = load i32, i32* %"&num_vs_constants[]"
>  store <8 x float> zeroinitializer, <8 x float>* %temp
>  store <8 x float> zeroinitializer, <8 x float>* %temp8
>  store <8 x float> zeroinitializer, <8 x float>* %temp9
>  store <8 x float> zeroinitializer, <8 x float>* %temp10
>  %22 = getelementptr float, float* %"vs_constants[]", i32 0
>  %23 = load float, float* %22
>  %24 = insertelement <8 x float> undef, float %23, i32 0
>  %25 = shufflevector <8 x float> %24, <8 x float> undef, <8 x i32> zeroinitializer
>  %26 = fmul <8 x float> %3, %25
>  %27 = getelementptr float, float* %"vs_constants[]", i32 1
>  %28 = load float, float* %27
>  %29 = insertelement <8 x float> undef, float %28, i32 0
>  %30 = shufflevector <8 x float> %29, <8 x float> undef, <8 x i32> zeroinitializer
>  %31 = fmul <8 x float> %3, %30
>  %32 = getelementptr float, float* %"vs_constants[]", i32 2
>  %33 = load float, float* %32
>  %34 = insertelement <8 x float> undef, float %33, i32 0
>  %35 = shufflevector <8 x float> %34, <8 x float> undef, <8 x i32> zeroinitializer
>  %36 = fmul <8 x float> %3, %35
>  %37 = getelementptr float, float* %"vs_constants[]", i32 3
>  %38 = load float, float* %37
>  %39 = insertelement <8 x float> undef, float %38, i32 0
>  %40 = shufflevector <8 x float> %39, <8 x float> undef, <8 x i32> zeroinitializer
>  %41 = fmul <8 x float> %3, %40
>  store <8 x float> %26, <8 x float>* %temp
>  store <8 x float> %31, <8 x float>* %temp8
>  store <8 x float> %36, <8 x float>* %temp9
>  store <8 x float> %41, <8 x float>* %temp10
>  %42 = getelementptr float, float* %"vs_constants[]", i32 4
>  %43 = load float, float* %42
>  %44 = insertelement <8 x float> undef, float %43, i32 0
>  %45 = shufflevector <8 x float> %44, <8 x float> undef, <8 x i32> zeroinitializer
>  %46 = load <8 x float>, <8 x float>* %temp
>  %47 = fmul <8 x float> %5, %45
>  %48 = fadd <8 x float> %47, %46
>  %49 = getelementptr float, float* %"vs_constants[]", i32 5
>  %50 = load float, float* %49
>  %51 = insertelement <8 x float> undef, float %50, i32 0
>  %52 = shufflevector <8 x float> %51, <8 x float> undef, <8 x i32> zeroinitializer
>  %53 = load <8 x float>, <8 x float>* %temp8
>  %54 = fmul <8 x float> %5, %52
>  %55 = fadd <8 x float> %54, %53
>  %56 = getelementptr float, float* %"vs_constants[]", i32 6
>  %57 = load float, float* %56
>  %58 = insertelement <8 x float> undef, float %57, i32 0
>  %59 = shufflevector <8 x float> %58, <8 x float> undef, <8 x i32> zeroinitializer
>  %60 = load <8 x float>, <8 x float>* %temp9
>  %61 = fmul <8 x float> %5, %59
>  %62 = fadd <8 x float> %61, %60
>  %63 = getelementptr float, float* %"vs_constants[]", i32 7
>  %64 = load float, float* %63
>  %65 = insertelement <8 x float> undef, float %64, i32 0
>  %66 = shufflevector <8 x float> %65, <8 x float> undef, <8 x i32> zeroinitializer
>  %67 = load <8 x float>, <8 x float>* %temp10
>  %68 = fmul <8 x float> %5, %66
>  %69 = fadd <8 x float> %68, %67
>  store <8 x float> %48, <8 x float>* %temp
>  store <8 x float> %55, <8 x float>* %temp8
>  store <8 x float> %62, <8 x float>* %temp9
>  store <8 x float> %69, <8 x float>* %temp10
>  %70 = getelementptr float, float* %"vs_constants[]", i32 8
>  %71 = load float, float* %70
>  %72 = insertelement <8 x float> undef, float %71, i32 0
>  %73 = shufflevector <8 x float> %72, <8 x float> undef, <8 x i32> zeroinitializer
>  %74 = load <8 x float>, <8 x float>* %temp
>  %75 = fmul <8 x float> %7, %73
>  %76 = fadd <8 x float> %75, %74
>  %77 = getelementptr float, float* %"vs_constants[]", i32 9
>  %78 = load float, float* %77
>  %79 = insertelement <8 x float> undef, float %78, i32 0
>  %80 = shufflevector <8 x float> %79, <8 x float> undef, <8 x i32> zeroinitializer
>  %81 = load <8 x float>, <8 x float>* %temp8
>  %82 = fmul <8 x float> %7, %80
>  %83 = fadd <8 x float> %82, %81
>  %84 = getelementptr float, float* %"vs_constants[]", i32 10
>  %85 = load float, float* %84
>  %86 = insertelement <8 x float> undef, float %85, i32 0
>  %87 = shufflevector <8 x float> %86, <8 x float> undef, <8 x i32> zeroinitializer
>  %88 = load <8 x float>, <8 x float>* %temp9
>  %89 = fmul <8 x float> %7, %87
>  %90 = fadd <8 x float> %89, %88
>  %91 = getelementptr float, float* %"vs_constants[]", i32 11
>  %92 = load float, float* %91
>  %93 = insertelement <8 x float> undef, float %92, i32 0
>  %94 = shufflevector <8 x float> %93, <8 x float> undef, <8 x i32> zeroinitializer
>  %95 = load <8 x float>, <8 x float>* %temp10
>  %96 = fmul <8 x float> %7, %94
>  %97 = fadd <8 x float> %96, %95
>  store <8 x float> %76, <8 x float>* %temp
>  store <8 x float> %83, <8 x float>* %temp8
>  store <8 x float> %90, <8 x float>* %temp9
>  store <8 x float> %97, <8 x float>* %temp10
>  %98 = getelementptr float, float* %"vs_constants[]", i32 12
>  %99 = load float, float* %98
>  %100 = insertelement <8 x float> undef, float %99, i32 0
>  %101 = shufflevector <8 x float> %100, <8 x float> undef, <8 x i32> zeroinitializer
>  %102 = load <8 x float>, <8 x float>* %temp
>  %103 = fmul <8 x float> %9, %101
>  %104 = fadd <8 x float> %103, %102
>  %105 = getelementptr float, float* %"vs_constants[]", i32 13
>  %106 = load float, float* %105
>  %107 = insertelement <8 x float> undef, float %106, i32 0
>  %108 = shufflevector <8 x float> %107, <8 x float> undef, <8 x i32> zeroinitializer
>  %109 = load <8 x float>, <8 x float>* %temp8
>  %110 = fmul <8 x float> %9, %108
>  %111 = fadd <8 x float> %110, %109
>  %112 = getelementptr float, float* %"vs_constants[]", i32 14
>  %113 = load float, float* %112
>  %114 = insertelement <8 x float> undef, float %113, i32 0
>  %115 = shufflevector <8 x float> %114, <8 x float> undef, <8 x i32> zeroinitializer
>  %116 = load <8 x float>, <8 x float>* %temp9
>  %117 = fmul <8 x float> %9, %115
>  %118 = fadd <8 x float> %117, %116
>  %119 = getelementptr float, float* %"vs_constants[]", i32 15
>  %120 = load float, float* %119
>  %121 = insertelement <8 x float> undef, float %120, i32 0
>  %122 = shufflevector <8 x float> %121, <8 x float> undef, <8 x i32> zeroinitializer
>  %123 = load <8 x float>, <8 x float>* %temp10
>  %124 = fmul <8 x float> %9, %122
>  %125 = fadd <8 x float> %124, %123
>  store <8 x float> %104, <8 x float>* %output
>  store <8 x float> %111, <8 x float>* %output1
>  store <8 x float> %118, <8 x float>* %output2
>  store <8 x float> %125, <8 x float>* %output3
>  store <8 x float> %11, <8 x float>* %output4
>  store <8 x float> %13, <8 x float>* %output5
>  store <8 x float> %15, <8 x float>* %output6
>  store <8 x float> %17, <8 x float>* %output7
>  %126 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 1
>  %127 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %126
>  %128 = load <8 x float>, <8 x float>* %output
>  %129 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 0, i32 0
>  store <8 x float> %128, <8 x float>* %129
>  %130 = load <8 x float>, <8 x float>* %output4
>  %131 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 1, i32 0
>  store <8 x float> %130, <8 x float>* %131
>  %132 = load <8 x float>, <8 x float>* %output1
>  %133 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 0, i32 1
>  store <8 x float> %132, <8 x float>* %133
>  %134 = load <8 x float>, <8 x float>* %output5
>  %135 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 1, i32 1
>  store <8 x float> %134, <8 x float>* %135
>  %136 = load <8 x float>, <8 x float>* %output2
>  %137 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 0, i32 2
>  store <8 x float> %136, <8 x float>* %137
>  %138 = load <8 x float>, <8 x float>* %output6
>  %139 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 1, i32 2
>  store <8 x float> %138, <8 x float>* %139
>  %140 = load <8 x float>, <8 x float>* %output3
>  %141 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 0, i32 3
>  store <8 x float> %140, <8 x float>* %141
>  %142 = load <8 x float>, <8 x float>* %output7
>  %143 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %127, i32 0, i32 0, i32 1, i32 3
>  store <8 x float> %142, <8 x float>* %143
>  ret void
>}
>
>VS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		andq	$-32, %rsp
>     8:		subq	$32, %rsp
>    12:		movq	(%rsi), %rcx
>    15:		movq	8(%rsi), %rax
>    19:		vmovaps	(%rcx), %ymm6
>    23:		vmovaps	32(%rcx), %ymm7
>    28:		vmovaps	64(%rcx), %ymm5
>    33:		vmovaps	96(%rcx), %ymm4
>    38:		vmovaps	128(%rcx), %ymm3
>    46:		vmovaps	160(%rcx), %ymm2
>    54:		vmovaps	192(%rcx), %ymm1
>    62:		vmovaps	224(%rcx), %ymm0
>    70:		movq	(%rdi), %rcx
>    73:		vbroadcastss	(%rcx), %ymm8
>    78:		vmulps	%ymm8, %ymm6, %ymm8
>    83:		vbroadcastss	4(%rcx), %ymm9
>    89:		vmulps	%ymm9, %ymm6, %ymm9
>    94:		vbroadcastss	8(%rcx), %ymm10
>   100:		vmulps	%ymm10, %ymm6, %ymm10
>   105:		vbroadcastss	12(%rcx), %ymm11
>   111:		vmulps	%ymm11, %ymm6, %ymm6
>   116:		vbroadcastss	16(%rcx), %ymm11
>   122:		vmulps	%ymm11, %ymm7, %ymm11
>   127:		vaddps	%ymm11, %ymm8, %ymm8
>   132:		vbroadcastss	20(%rcx), %ymm11
>   138:		vmulps	%ymm11, %ymm7, %ymm11
>   143:		vaddps	%ymm11, %ymm9, %ymm9
>   148:		vbroadcastss	24(%rcx), %ymm11
>   154:		vmulps	%ymm11, %ymm7, %ymm11
>   159:		vaddps	%ymm11, %ymm10, %ymm10
>   164:		vbroadcastss	28(%rcx), %ymm11
>   170:		vmulps	%ymm11, %ymm7, %ymm7
>   175:		vaddps	%ymm7, %ymm6, %ymm6
>   179:		vbroadcastss	32(%rcx), %ymm7
>   185:		vmulps	%ymm7, %ymm5, %ymm7
>   189:		vaddps	%ymm7, %ymm8, %ymm7
>   193:		vbroadcastss	36(%rcx), %ymm8
>   199:		vmulps	%ymm8, %ymm5, %ymm8
>   204:		vaddps	%ymm8, %ymm9, %ymm8
>   209:		vbroadcastss	40(%rcx), %ymm9
>   215:		vmulps	%ymm9, %ymm5, %ymm9
>   220:		vaddps	%ymm9, %ymm10, %ymm9
>   225:		vbroadcastss	44(%rcx), %ymm10
>   231:		vmulps	%ymm10, %ymm5, %ymm5
>   236:		vaddps	%ymm5, %ymm6, %ymm5
>   240:		vbroadcastss	48(%rcx), %ymm6
>   246:		vmulps	%ymm6, %ymm4, %ymm6
>   250:		vaddps	%ymm6, %ymm7, %ymm6
>   254:		vbroadcastss	52(%rcx), %ymm7
>   260:		vmulps	%ymm7, %ymm4, %ymm7
>   264:		vaddps	%ymm7, %ymm8, %ymm7
>   268:		vbroadcastss	56(%rcx), %ymm8
>   274:		vmulps	%ymm8, %ymm4, %ymm8
>   279:		vaddps	%ymm8, %ymm9, %ymm8
>   284:		vbroadcastss	60(%rcx), %ymm9
>   290:		vmulps	%ymm9, %ymm4, %ymm4
>   295:		vaddps	%ymm4, %ymm5, %ymm4
>   299:		vmovaps	%ymm6, (%rax)
>   303:		vmovaps	%ymm3, 128(%rax)
>   311:		vmovaps	%ymm7, 32(%rax)
>   316:		vmovaps	%ymm2, 160(%rax)
>   324:		vmovaps	%ymm8, 64(%rax)
>   329:		vmovaps	%ymm1, 192(%rax)
>   337:		vmovaps	%ymm4, 96(%rax)
>   342:		vmovaps	%ymm0, 224(%rax)
>   350:		movq	%rbp, %rsp
>   353:		popq	%rbp
>   354:		vzeroupper
>   357:		retq
>
>
>vert shader  0x7f7aa462f000
>
>; Function Attrs: alignstack(32)
>define void @FS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx) #0 {
>entry:
>  %output25 = alloca <8 x float>
>  %output24 = alloca <8 x float>
>  %output23 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 2
>  %num_fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 3
>  %0 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 9
>  %pRawAttribs = load float*, float** %0
>  %1 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 10
>  %pPerspAttribs = load float*, float** %1
>  %2 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i = load <8 x float>, <8 x float>* %2
>  %3 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j = load <8 x float>, <8 x float>* %3
>  %4 = getelementptr float, float* %pRawAttribs, i32 0
>  %5 = load float, float* %4
>  %.splatinsert = insertelement <8 x float> undef, float %5, i32 0
>  %.splat = shufflevector <8 x float> %.splatinsert, <8 x float> undef, <8 x i32> zeroinitializer
>  %6 = getelementptr float, float* %pRawAttribs, i32 4
>  %7 = load float, float* %6
>  %.splatinsert1 = insertelement <8 x float> undef, float %7, i32 0
>  %.splat2 = shufflevector <8 x float> %.splatinsert1, <8 x float> undef, <8 x i32> zeroinitializer
>  %8 = getelementptr float, float* %pRawAttribs, i32 8
>  %9 = load float, float* %8
>  %.splatinsert3 = insertelement <8 x float> undef, float %9, i32 0
>  %.splat4 = shufflevector <8 x float> %.splatinsert3, <8 x float> undef, <8 x i32> zeroinitializer
>  %10 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %11 = fsub <8 x float> %10, %j
>  %12 = fmul <8 x float> %11, %.splat4
>  %13 = fmul <8 x float> %.splat, %i
>  %14 = fmul <8 x float> %.splat2, %j
>  %15 = fadd <8 x float> %13, %14
>  %16 = fadd <8 x float> %15, %12
>  %17 = getelementptr float, float* %pRawAttribs, i32 1
>  %18 = load float, float* %17
>  %.splatinsert5 = insertelement <8 x float> undef, float %18, i32 0
>  %.splat6 = shufflevector <8 x float> %.splatinsert5, <8 x float> undef, <8 x i32> zeroinitializer
>  %19 = getelementptr float, float* %pRawAttribs, i32 5
>  %20 = load float, float* %19
>  %.splatinsert7 = insertelement <8 x float> undef, float %20, i32 0
>  %.splat8 = shufflevector <8 x float> %.splatinsert7, <8 x float> undef, <8 x i32> zeroinitializer
>  %21 = getelementptr float, float* %pRawAttribs, i32 9
>  %22 = load float, float* %21
>  %.splatinsert9 = insertelement <8 x float> undef, float %22, i32 0
>  %.splat10 = shufflevector <8 x float> %.splatinsert9, <8 x float> undef, <8 x i32> zeroinitializer
>  %23 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %24 = fsub <8 x float> %23, %j
>  %25 = fmul <8 x float> %24, %.splat10
>  %26 = fmul <8 x float> %.splat6, %i
>  %27 = fmul <8 x float> %.splat8, %j
>  %28 = fadd <8 x float> %26, %27
>  %29 = fadd <8 x float> %28, %25
>  %30 = getelementptr float, float* %pRawAttribs, i32 2
>  %31 = load float, float* %30
>  %.splatinsert11 = insertelement <8 x float> undef, float %31, i32 0
>  %.splat12 = shufflevector <8 x float> %.splatinsert11, <8 x float> undef, <8 x i32> zeroinitializer
>  %32 = getelementptr float, float* %pRawAttribs, i32 6
>  %33 = load float, float* %32
>  %.splatinsert13 = insertelement <8 x float> undef, float %33, i32 0
>  %.splat14 = shufflevector <8 x float> %.splatinsert13, <8 x float> undef, <8 x i32> zeroinitializer
>  %34 = getelementptr float, float* %pRawAttribs, i32 10
>  %35 = load float, float* %34
>  %.splatinsert15 = insertelement <8 x float> undef, float %35, i32 0
>  %.splat16 = shufflevector <8 x float> %.splatinsert15, <8 x float> undef, <8 x i32> zeroinitializer
>  %36 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %37 = fsub <8 x float> %36, %j
>  %38 = fmul <8 x float> %37, %.splat16
>  %39 = fmul <8 x float> %.splat12, %i
>  %40 = fmul <8 x float> %.splat14, %j
>  %41 = fadd <8 x float> %39, %40
>  %42 = fadd <8 x float> %41, %38
>  %43 = getelementptr float, float* %pRawAttribs, i32 3
>  %44 = load float, float* %43
>  %.splatinsert17 = insertelement <8 x float> undef, float %44, i32 0
>  %.splat18 = shufflevector <8 x float> %.splatinsert17, <8 x float> undef, <8 x i32> zeroinitializer
>  %45 = getelementptr float, float* %pRawAttribs, i32 7
>  %46 = load float, float* %45
>  %.splatinsert19 = insertelement <8 x float> undef, float %46, i32 0
>  %.splat20 = shufflevector <8 x float> %.splatinsert19, <8 x float> undef, <8 x i32> zeroinitializer
>  %47 = getelementptr float, float* %pRawAttribs, i32 11
>  %48 = load float, float* %47
>  %.splatinsert21 = insertelement <8 x float> undef, float %48, i32 0
>  %.splat22 = shufflevector <8 x float> %.splatinsert21, <8 x float> undef, <8 x i32> zeroinitializer
>  %49 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %50 = fsub <8 x float> %49, %j
>  %51 = fmul <8 x float> %50, %.splat22
>  %52 = fmul <8 x float> %.splat18, %i
>  %53 = fmul <8 x float> %.splat20, %j
>  %54 = fadd <8 x float> %52, %53
>  %55 = fadd <8 x float> %54, %51
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output23
>  store <8 x float> zeroinitializer, <8 x float>* %output24
>  store <8 x float> zeroinitializer, <8 x float>* %output25
>  store <8 x float> %16, <8 x float>* %output
>  store <8 x float> %29, <8 x float>* %output23
>  store <8 x float> %42, <8 x float>* %output24
>  store <8 x float> %55, <8 x float>* %output25
>  %56 = load <8 x float>, <8 x float>* %output
>  %57 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 0
>  store <8 x float> %56, <8 x float>* %57
>  %58 = load <8 x float>, <8 x float>* %output23
>  %59 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 1
>  store <8 x float> %58, <8 x float>* %59
>  %60 = load <8 x float>, <8 x float>* %output24
>  %61 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 2
>  store <8 x float> %60, <8 x float>* %61
>  %62 = load <8 x float>, <8 x float>* %output25
>  %63 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 3
>  store <8 x float> %62, <8 x float>* %63
>  ret void
>}
>
>FS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		andq	$-32, %rsp
>     8:		subq	$32, %rsp
>    12:		movq	768(%rsi), %rax
>    19:		vmovaps	416(%rsi), %ymm0
>    27:		vmovaps	544(%rsi), %ymm1
>    35:		vbroadcastss	(%rax), %ymm2
>    40:		vbroadcastss	16(%rax), %ymm3
>    46:		vbroadcastss	32(%rax), %ymm4
>    52:		movabsq	$140164720672768, %rcx
>    62:		vbroadcastss	(%rcx), %ymm5
>    67:		vsubps	%ymm0, %ymm5, %ymm5
>    71:		vsubps	%ymm1, %ymm5, %ymm5
>    75:		vmulps	%ymm4, %ymm5, %ymm4
>    79:		vmulps	%ymm2, %ymm0, %ymm2
>    83:		vmulps	%ymm3, %ymm1, %ymm3
>    87:		vaddps	%ymm3, %ymm2, %ymm2
>    91:		vaddps	%ymm4, %ymm2, %ymm2
>    95:		vbroadcastss	4(%rax), %ymm3
>   101:		vbroadcastss	20(%rax), %ymm4
>   107:		vbroadcastss	36(%rax), %ymm6
>   113:		vmulps	%ymm6, %ymm5, %ymm6
>   117:		vmulps	%ymm3, %ymm0, %ymm3
>   121:		vmulps	%ymm4, %ymm1, %ymm4
>   125:		vaddps	%ymm4, %ymm3, %ymm3
>   129:		vaddps	%ymm6, %ymm3, %ymm3
>   133:		vbroadcastss	8(%rax), %ymm4
>   139:		vbroadcastss	24(%rax), %ymm6
>   145:		vbroadcastss	40(%rax), %ymm7
>   151:		vmulps	%ymm7, %ymm5, %ymm7
>   155:		vmulps	%ymm4, %ymm0, %ymm4
>   159:		vmulps	%ymm6, %ymm1, %ymm6
>   163:		vaddps	%ymm6, %ymm4, %ymm4
>   167:		vaddps	%ymm7, %ymm4, %ymm4
>   171:		vbroadcastss	12(%rax), %ymm6
>   177:		vbroadcastss	28(%rax), %ymm7
>   183:		vbroadcastss	44(%rax), %ymm8
>   189:		vmulps	%ymm8, %ymm5, %ymm5
>   194:		vmulps	%ymm6, %ymm0, %ymm0
>   198:		vmulps	%ymm7, %ymm1, %ymm1
>   202:		vaddps	%ymm1, %ymm0, %ymm0
>   206:		vaddps	%ymm5, %ymm0, %ymm0
>   210:		vmovaps	%ymm2, 832(%rsi)
>   218:		vmovaps	%ymm3, 864(%rsi)
>   226:		vmovaps	%ymm4, 896(%rsi)
>   234:		vmovaps	%ymm0, 928(%rsi)
>   242:		movq	%rbp, %rsp
>   245:		popq	%rbp
>   246:		vzeroupper
>   249:		retq
>
>
>frag shader  0x7f7aa462d000
>BLEND shader 0x7f7aa462b000
>dual depth peeling
>FRAG
>PROPERTY FS_COLOR0_WRITES_ALL_CBUFS 1
>DCL IN[0], COLOR, COLOR
>DCL OUT[0], COLOR
>  0: MOV OUT[0], IN[0]
>  1: END
>OUT[0].x = IN[0].x
>OUT[0].y = IN[0].y
>OUT[0].z = IN[0].z
>OUT[0].w = IN[0].w
>
>; Function Attrs: alignstack(32)
>define void @FS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx) #0 {
>entry:
>  %output25 = alloca <8 x float>
>  %output24 = alloca <8 x float>
>  %output23 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 2
>  %num_fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 3
>  %0 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 9
>  %pRawAttribs = load float*, float** %0
>  %1 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 10
>  %pPerspAttribs = load float*, float** %1
>  %2 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i = load <8 x float>, <8 x float>* %2
>  %3 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j = load <8 x float>, <8 x float>* %3
>  %4 = getelementptr float, float* %pRawAttribs, i32 0
>  %5 = load float, float* %4
>  %.splatinsert = insertelement <8 x float> undef, float %5, i32 0
>  %.splat = shufflevector <8 x float> %.splatinsert, <8 x float> undef, <8 x i32> zeroinitializer
>  %6 = getelementptr float, float* %pRawAttribs, i32 4
>  %7 = load float, float* %6
>  %.splatinsert1 = insertelement <8 x float> undef, float %7, i32 0
>  %.splat2 = shufflevector <8 x float> %.splatinsert1, <8 x float> undef, <8 x i32> zeroinitializer
>  %8 = getelementptr float, float* %pRawAttribs, i32 8
>  %9 = load float, float* %8
>  %.splatinsert3 = insertelement <8 x float> undef, float %9, i32 0
>  %.splat4 = shufflevector <8 x float> %.splatinsert3, <8 x float> undef, <8 x i32> zeroinitializer
>  %10 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %11 = fsub <8 x float> %10, %j
>  %12 = fmul <8 x float> %11, %.splat4
>  %13 = fmul <8 x float> %.splat, %i
>  %14 = fmul <8 x float> %.splat2, %j
>  %15 = fadd <8 x float> %13, %14
>  %16 = fadd <8 x float> %15, %12
>  %17 = getelementptr float, float* %pRawAttribs, i32 1
>  %18 = load float, float* %17
>  %.splatinsert5 = insertelement <8 x float> undef, float %18, i32 0
>  %.splat6 = shufflevector <8 x float> %.splatinsert5, <8 x float> undef, <8 x i32> zeroinitializer
>  %19 = getelementptr float, float* %pRawAttribs, i32 5
>  %20 = load float, float* %19
>  %.splatinsert7 = insertelement <8 x float> undef, float %20, i32 0
>  %.splat8 = shufflevector <8 x float> %.splatinsert7, <8 x float> undef, <8 x i32> zeroinitializer
>  %21 = getelementptr float, float* %pRawAttribs, i32 9
>  %22 = load float, float* %21
>  %.splatinsert9 = insertelement <8 x float> undef, float %22, i32 0
>  %.splat10 = shufflevector <8 x float> %.splatinsert9, <8 x float> undef, <8 x i32> zeroinitializer
>  %23 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %24 = fsub <8 x float> %23, %j
>  %25 = fmul <8 x float> %24, %.splat10
>  %26 = fmul <8 x float> %.splat6, %i
>  %27 = fmul <8 x float> %.splat8, %j
>  %28 = fadd <8 x float> %26, %27
>  %29 = fadd <8 x float> %28, %25
>  %30 = getelementptr float, float* %pRawAttribs, i32 2
>  %31 = load float, float* %30
>  %.splatinsert11 = insertelement <8 x float> undef, float %31, i32 0
>  %.splat12 = shufflevector <8 x float> %.splatinsert11, <8 x float> undef, <8 x i32> zeroinitializer
>  %32 = getelementptr float, float* %pRawAttribs, i32 6
>  %33 = load float, float* %32
>  %.splatinsert13 = insertelement <8 x float> undef, float %33, i32 0
>  %.splat14 = shufflevector <8 x float> %.splatinsert13, <8 x float> undef, <8 x i32> zeroinitializer
>  %34 = getelementptr float, float* %pRawAttribs, i32 10
>  %35 = load float, float* %34
>  %.splatinsert15 = insertelement <8 x float> undef, float %35, i32 0
>  %.splat16 = shufflevector <8 x float> %.splatinsert15, <8 x float> undef, <8 x i32> zeroinitializer
>  %36 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %37 = fsub <8 x float> %36, %j
>  %38 = fmul <8 x float> %37, %.splat16
>  %39 = fmul <8 x float> %.splat12, %i
>  %40 = fmul <8 x float> %.splat14, %j
>  %41 = fadd <8 x float> %39, %40
>  %42 = fadd <8 x float> %41, %38
>  %43 = getelementptr float, float* %pRawAttribs, i32 3
>  %44 = load float, float* %43
>  %.splatinsert17 = insertelement <8 x float> undef, float %44, i32 0
>  %.splat18 = shufflevector <8 x float> %.splatinsert17, <8 x float> undef, <8 x i32> zeroinitializer
>  %45 = getelementptr float, float* %pRawAttribs, i32 7
>  %46 = load float, float* %45
>  %.splatinsert19 = insertelement <8 x float> undef, float %46, i32 0
>  %.splat20 = shufflevector <8 x float> %.splatinsert19, <8 x float> undef, <8 x i32> zeroinitializer
>  %47 = getelementptr float, float* %pRawAttribs, i32 11
>  %48 = load float, float* %47
>  %.splatinsert21 = insertelement <8 x float> undef, float %48, i32 0
>  %.splat22 = shufflevector <8 x float> %.splatinsert21, <8 x float> undef, <8 x i32> zeroinitializer
>  %49 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %50 = fsub <8 x float> %49, %j
>  %51 = fmul <8 x float> %50, %.splat22
>  %52 = fmul <8 x float> %.splat18, %i
>  %53 = fmul <8 x float> %.splat20, %j
>  %54 = fadd <8 x float> %52, %53
>  %55 = fadd <8 x float> %54, %51
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output23
>  store <8 x float> zeroinitializer, <8 x float>* %output24
>  store <8 x float> zeroinitializer, <8 x float>* %output25
>  store <8 x float> %16, <8 x float>* %output
>  store <8 x float> %29, <8 x float>* %output23
>  store <8 x float> %42, <8 x float>* %output24
>  store <8 x float> %55, <8 x float>* %output25
>  %56 = load <8 x float>, <8 x float>* %output
>  %57 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 0
>  store <8 x float> %56, <8 x float>* %57
>  %58 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 0
>  store <8 x float> %56, <8 x float>* %58
>  %59 = load <8 x float>, <8 x float>* %output23
>  %60 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 1
>  store <8 x float> %59, <8 x float>* %60
>  %61 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 1
>  store <8 x float> %59, <8 x float>* %61
>  %62 = load <8 x float>, <8 x float>* %output24
>  %63 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 2
>  store <8 x float> %62, <8 x float>* %63
>  %64 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 2
>  store <8 x float> %62, <8 x float>* %64
>  %65 = load <8 x float>, <8 x float>* %output25
>  %66 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 3
>  store <8 x float> %65, <8 x float>* %66
>  %67 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 3
>  store <8 x float> %65, <8 x float>* %67
>  ret void
>}
>
>FS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		andq	$-32, %rsp
>     8:		subq	$32, %rsp
>    12:		movq	768(%rsi), %rax
>    19:		vmovaps	416(%rsi), %ymm0
>    27:		vmovaps	544(%rsi), %ymm1
>    35:		vbroadcastss	(%rax), %ymm2
>    40:		vbroadcastss	16(%rax), %ymm3
>    46:		vbroadcastss	32(%rax), %ymm4
>    52:		movabsq	$140164720652288, %rcx
>    62:		vbroadcastss	(%rcx), %ymm5
>    67:		vsubps	%ymm0, %ymm5, %ymm5
>    71:		vsubps	%ymm1, %ymm5, %ymm5
>    75:		vmulps	%ymm4, %ymm5, %ymm4
>    79:		vmulps	%ymm2, %ymm0, %ymm2
>    83:		vmulps	%ymm3, %ymm1, %ymm3
>    87:		vaddps	%ymm3, %ymm2, %ymm2
>    91:		vaddps	%ymm4, %ymm2, %ymm2
>    95:		vbroadcastss	4(%rax), %ymm3
>   101:		vbroadcastss	20(%rax), %ymm4
>   107:		vbroadcastss	36(%rax), %ymm6
>   113:		vmulps	%ymm6, %ymm5, %ymm6
>   117:		vmulps	%ymm3, %ymm0, %ymm3
>   121:		vmulps	%ymm4, %ymm1, %ymm4
>   125:		vaddps	%ymm4, %ymm3, %ymm3
>   129:		vaddps	%ymm6, %ymm3, %ymm3
>   133:		vbroadcastss	8(%rax), %ymm4
>   139:		vbroadcastss	24(%rax), %ymm6
>   145:		vbroadcastss	40(%rax), %ymm7
>   151:		vmulps	%ymm7, %ymm5, %ymm7
>   155:		vmulps	%ymm4, %ymm0, %ymm4
>   159:		vmulps	%ymm6, %ymm1, %ymm6
>   163:		vaddps	%ymm6, %ymm4, %ymm4
>   167:		vaddps	%ymm7, %ymm4, %ymm4
>   171:		vbroadcastss	12(%rax), %ymm6
>   177:		vbroadcastss	28(%rax), %ymm7
>   183:		vbroadcastss	44(%rax), %ymm8
>   189:		vmulps	%ymm8, %ymm5, %ymm5
>   194:		vmulps	%ymm6, %ymm0, %ymm0
>   198:		vmulps	%ymm7, %ymm1, %ymm1
>   202:		vaddps	%ymm1, %ymm0, %ymm0
>   206:		vaddps	%ymm5, %ymm0, %ymm0
>   210:		vmovaps	%ymm2, 832(%rsi)
>   218:		vmovaps	%ymm2, 960(%rsi)
>   226:		vmovaps	%ymm3, 864(%rsi)
>   234:		vmovaps	%ymm3, 992(%rsi)
>   242:		vmovaps	%ymm4, 896(%rsi)
>   250:		vmovaps	%ymm4, 1024(%rsi)
>   258:		vmovaps	%ymm0, 928(%rsi)
>   266:		vmovaps	%ymm0, 1056(%rsi)
>   274:		movq	%rbp, %rsp
>   277:		popq	%rbp
>   278:		vzeroupper
>   281:		retq
>
>
>frag shader  0x7f7aa4628000
>BLEND shader 0x7f7aa4626000
>BLEND shader 0x7f7aa4624000
>FRAG
>DCL IN[0], GENERIC[0], LINEAR
>DCL OUT[0], COLOR
>DCL OUT[1], POSITION
>DCL SAMP[0]
>DCL SVIEW[0], 2D, FLOAT
>IMM[0] FLT32 {    0.0000,     1.0000,     0.0000,     0.0000}
>  0: MOV OUT[0], IMM[0].xxxy
>  1: TEX OUT[1].z, IN[0], SAMP[0], 2D
>  2: END
>TEX[0] = IN[0].x IN[0].y _ _, RES[0], SAMP[0], 2D
>OUT[0].x = 0.000000
>OUT[0].y = 0.000000
>OUT[0].z = 0.000000
>OUT[0].w = 1.000000
>VERT
>PROPERTY NEXT_SHADER 0
>DCL IN[0]
>DCL IN[1]
>DCL OUT[0], POSITION
>DCL OUT[1], GENERIC[0]
>  0: MOV OUT[0], IN[0]
>  1: MOV OUT[1], IN[1]
>  2: END
>OUT[0].x = IN[0].x
>OUT[0].y = IN[0].y
>OUT[0].z = IN[0].z
>OUT[0].w = IN[0].w
>OUT[1].x = IN[1].x
>OUT[1].y = IN[1].y
>OUT[1].z = IN[1].z
>OUT[1].w = IN[1].w
>
>; Function Attrs: alignstack(32)
>define void @VS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx) #0 {
>entry:
>  %output7 = alloca <8 x float>
>  %output6 = alloca <8 x float>
>  %output5 = alloca <8 x float>
>  %output4 = alloca <8 x float>
>  %output3 = alloca <8 x float>
>  %output2 = alloca <8 x float>
>  %output1 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 0
>  %num_vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 1
>  %0 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 0
>  %1 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %0
>  %2 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 0
>  %3 = load <8 x float>, <8 x float>* %2
>  %4 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 1
>  %5 = load <8 x float>, <8 x float>* %4
>  %6 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 2
>  %7 = load <8 x float>, <8 x float>* %6
>  %8 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 3
>  %9 = load <8 x float>, <8 x float>* %8
>  %10 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 0
>  %11 = load <8 x float>, <8 x float>* %10
>  %12 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 1
>  %13 = load <8 x float>, <8 x float>* %12
>  %14 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 2
>  %15 = load <8 x float>, <8 x float>* %14
>  %16 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 3
>  %17 = load <8 x float>, <8 x float>* %16
>  %18 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 2
>  %19 = load i32, i32* %18
>  %20 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 3
>  %21 = load <8 x i32>, <8 x i32>* %20
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output1
>  store <8 x float> zeroinitializer, <8 x float>* %output2
>  store <8 x float> zeroinitializer, <8 x float>* %output3
>  store <8 x float> zeroinitializer, <8 x float>* %output4
>  store <8 x float> zeroinitializer, <8 x float>* %output5
>  store <8 x float> zeroinitializer, <8 x float>* %output6
>  store <8 x float> zeroinitializer, <8 x float>* %output7
>  store <8 x float> %3, <8 x float>* %output
>  store <8 x float> %5, <8 x float>* %output1
>  store <8 x float> %7, <8 x float>* %output2
>  store <8 x float> %9, <8 x float>* %output3
>  store <8 x float> %11, <8 x float>* %output4
>  store <8 x float> %13, <8 x float>* %output5
>  store <8 x float> %15, <8 x float>* %output6
>  store <8 x float> %17, <8 x float>* %output7
>  %22 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 1
>  %23 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %22
>  %24 = load <8 x float>, <8 x float>* %output
>  %25 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 0
>  store <8 x float> %24, <8 x float>* %25
>  %26 = load <8 x float>, <8 x float>* %output4
>  %27 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 0
>  store <8 x float> %26, <8 x float>* %27
>  %28 = load <8 x float>, <8 x float>* %output1
>  %29 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 1
>  store <8 x float> %28, <8 x float>* %29
>  %30 = load <8 x float>, <8 x float>* %output5
>  %31 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 1
>  store <8 x float> %30, <8 x float>* %31
>  %32 = load <8 x float>, <8 x float>* %output2
>  %33 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 2
>  store <8 x float> %32, <8 x float>* %33
>  %34 = load <8 x float>, <8 x float>* %output6
>  %35 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 2
>  store <8 x float> %34, <8 x float>* %35
>  %36 = load <8 x float>, <8 x float>* %output3
>  %37 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 3
>  store <8 x float> %36, <8 x float>* %37
>  %38 = load <8 x float>, <8 x float>* %output7
>  %39 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 3
>  store <8 x float> %38, <8 x float>* %39
>  ret void
>}
>
>VS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		andq	$-32, %rsp
>     8:		subq	$32, %rsp
>    12:		movq	(%rsi), %rax
>    15:		movq	8(%rsi), %rcx
>    19:		vmovaps	(%rax), %ymm0
>    23:		vmovaps	32(%rax), %ymm1
>    28:		vmovaps	64(%rax), %ymm2
>    33:		vmovaps	96(%rax), %ymm3
>    38:		vmovaps	128(%rax), %ymm4
>    46:		vmovaps	160(%rax), %ymm5
>    54:		vmovaps	192(%rax), %ymm6
>    62:		vmovaps	224(%rax), %ymm7
>    70:		vmovaps	%ymm0, (%rcx)
>    74:		vmovaps	%ymm4, 128(%rcx)
>    82:		vmovaps	%ymm1, 32(%rcx)
>    87:		vmovaps	%ymm5, 160(%rcx)
>    95:		vmovaps	%ymm2, 64(%rcx)
>   100:		vmovaps	%ymm6, 192(%rcx)
>   108:		vmovaps	%ymm3, 96(%rcx)
>   113:		vmovaps	%ymm7, 224(%rcx)
>   121:		movq	%rbp, %rsp
>   124:		popq	%rbp
>   125:		vzeroupper
>   128:		retq
>
>
>vert shader  0x7f7aa4622000
>
>; Function Attrs: alignstack(32)
>define void @FS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx) #0 {
>entry:
>  %sampler0_texel_w_var = alloca <8 x float>
>  %sampler0_texel_z_var = alloca <8 x float>
>  %sampler0_texel_y_var = alloca <8 x float>
>  %sampler0_texel_x_var = alloca <8 x float>
>  %output17 = alloca <8 x float>
>  %output16 = alloca <8 x float>
>  %output15 = alloca <8 x float>
>  %output14 = alloca <8 x float>
>  %output13 = alloca <8 x float>
>  %output12 = alloca <8 x float>
>  %output11 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 2
>  %num_fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 3
>  %0 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 9
>  %pRawAttribs = load float*, float** %0
>  %1 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 10
>  %pPerspAttribs = load float*, float** %1
>  %2 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i = load <8 x float>, <8 x float>* %2
>  %3 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j = load <8 x float>, <8 x float>* %3
>  %4 = getelementptr float, float* %pRawAttribs, i32 0
>  %5 = load float, float* %4
>  %.splatinsert = insertelement <8 x float> undef, float %5, i32 0
>  %.splat = shufflevector <8 x float> %.splatinsert, <8 x float> undef, <8 x i32> zeroinitializer
>  %6 = getelementptr float, float* %pRawAttribs, i32 4
>  %7 = load float, float* %6
>  %.splatinsert1 = insertelement <8 x float> undef, float %7, i32 0
>  %.splat2 = shufflevector <8 x float> %.splatinsert1, <8 x float> undef, <8 x i32> zeroinitializer
>  %8 = getelementptr float, float* %pRawAttribs, i32 8
>  %9 = load float, float* %8
>  %.splatinsert3 = insertelement <8 x float> undef, float %9, i32 0
>  %.splat4 = shufflevector <8 x float> %.splatinsert3, <8 x float> undef, <8 x i32> zeroinitializer
>  %10 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %11 = fsub <8 x float> %10, %j
>  %12 = fmul <8 x float> %11, %.splat4
>  %13 = fmul <8 x float> %.splat, %i
>  %14 = fmul <8 x float> %.splat2, %j
>  %15 = fadd <8 x float> %13, %14
>  %16 = fadd <8 x float> %15, %12
>  %17 = getelementptr float, float* %pRawAttribs, i32 1
>  %18 = load float, float* %17
>  %.splatinsert5 = insertelement <8 x float> undef, float %18, i32 0
>  %.splat6 = shufflevector <8 x float> %.splatinsert5, <8 x float> undef, <8 x i32> zeroinitializer
>  %19 = getelementptr float, float* %pRawAttribs, i32 5
>  %20 = load float, float* %19
>  %.splatinsert7 = insertelement <8 x float> undef, float %20, i32 0
>  %.splat8 = shufflevector <8 x float> %.splatinsert7, <8 x float> undef, <8 x i32> zeroinitializer
>  %21 = getelementptr float, float* %pRawAttribs, i32 9
>  %22 = load float, float* %21
>  %.splatinsert9 = insertelement <8 x float> undef, float %22, i32 0
>  %.splat10 = shufflevector <8 x float> %.splatinsert9, <8 x float> undef, <8 x i32> zeroinitializer
>  %23 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %24 = fsub <8 x float> %23, %j
>  %25 = fmul <8 x float> %24, %.splat10
>  %26 = fmul <8 x float> %.splat6, %i
>  %27 = fmul <8 x float> %.splat8, %j
>  %28 = fadd <8 x float> %26, %27
>  %29 = fadd <8 x float> %28, %25
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output11
>  store <8 x float> zeroinitializer, <8 x float>* %output12
>  store <8 x float> zeroinitializer, <8 x float>* %output13
>  store <8 x float> zeroinitializer, <8 x float>* %output14
>  store <8 x float> zeroinitializer, <8 x float>* %output15
>  store <8 x float> zeroinitializer, <8 x float>* %output16
>  store <8 x float> zeroinitializer, <8 x float>* %output17
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output11
>  store <8 x float> zeroinitializer, <8 x float>* %output12
>  store <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, <8 x float>* %output13
>  %30 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 0
>  %context.texture0.width = load i32, i32* %30
>  %context.texture0.row_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 6
>  %context.texture0.img_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 7
>  %31 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 5
>  %context.texture0.base_ptr = load i8*, i8** %31
>  %context.texture0.mip_offsets = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 8
>  %32 = insertelement <4 x i32> undef, i32 %context.texture0.width, i32 0
>  %33 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 1
>  %context.texture0.height = load i32, i32* %33
>  %34 = insertelement <4 x i32> %32, i32 %context.texture0.height, i32 1
>  %35 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 3
>  %context.texture0.first_level = load i32, i32* %35
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_x_var
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_y_var
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_z_var
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_w_var
>  %36 = insertelement <4 x i32> undef, i32 %context.texture0.first_level, i32 0
>  %37 = shufflevector <4 x i32> %36, <4 x i32> undef, <4 x i32> zeroinitializer
>  %minify = lshr <4 x i32> %34, %37
>  %38 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %39 = getelementptr [16 x i32], [16 x i32]* %context.texture0.row_stride, i32 0, i32 %context.texture0.first_level
>  %40 = load i32, i32* %39
>  %41 = insertelement <8 x i32> undef, i32 %40, i32 0
>  %42 = shufflevector <8 x i32> %41, <8 x i32> undef, <8 x i32> zeroinitializer
>  %43 = getelementptr [16 x i32], [16 x i32]* %context.texture0.mip_offsets, i32 0, i32 %context.texture0.first_level
>  %44 = load i32, i32* %43
>  %45 = getelementptr i8, i8* %context.texture0.base_ptr, i32 %44
>  %46 = shufflevector <4 x i32> %38, <4 x i32> undef, <8 x i32> zeroinitializer
>  %47 = shufflevector <4 x i32> %38, <4 x i32> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %48 = sitofp <4 x i32> %38 to <4 x float>
>  %49 = shufflevector <4 x float> %48, <4 x float> undef, <8 x i32> zeroinitializer
>  %50 = shufflevector <4 x float> %48, <4 x float> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %51 = sub <8 x i32> %46, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %52 = fmul <8 x float> %16, %49
>  %53 = fptosi <8 x float> %52 to <8 x i32>
>  %54 = shufflevector <8 x i32> %53, <8 x i32> %53, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %55 = shufflevector <8 x i32> %51, <8 x i32> %51, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %56 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %54, <4 x i32> %55)
>  %57 = shufflevector <8 x i32> %53, <8 x i32> %53, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %58 = shufflevector <8 x i32> %51, <8 x i32> %51, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %59 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %57, <4 x i32> %58)
>  %60 = shufflevector <4 x i32> %56, <4 x i32> %59, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %61 = shufflevector <8 x i32> %60, <8 x i32> %60, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %62 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %61, <4 x i32> zeroinitializer)
>  %63 = shufflevector <8 x i32> %60, <8 x i32> %60, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %64 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %63, <4 x i32> zeroinitializer)
>  %tex.x.wrapped = shufflevector <4 x i32> %62, <4 x i32> %64, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %65 = sub <8 x i32> %47, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %66 = fmul <8 x float> %29, %50
>  %67 = fptosi <8 x float> %66 to <8 x i32>
>  %68 = shufflevector <8 x i32> %67, <8 x i32> %67, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %69 = shufflevector <8 x i32> %65, <8 x i32> %65, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %70 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %68, <4 x i32> %69)
>  %71 = shufflevector <8 x i32> %67, <8 x i32> %67, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %72 = shufflevector <8 x i32> %65, <8 x i32> %65, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %73 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %71, <4 x i32> %72)
>  %74 = shufflevector <4 x i32> %70, <4 x i32> %73, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %75 = shufflevector <8 x i32> %74, <8 x i32> %74, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %76 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %75, <4 x i32> zeroinitializer)
>  %77 = shufflevector <8 x i32> %74, <8 x i32> %74, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %78 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %77, <4 x i32> zeroinitializer)
>  %tex.y.wrapped = shufflevector <4 x i32> %76, <4 x i32> %78, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %79 = mul <8 x i32> %tex.x.wrapped, <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4>
>  %80 = mul <8 x i32> %tex.y.wrapped, %42
>  %81 = add <8 x i32> %79, %80
>  %82 = extractelement <8 x i32> %81, i32 0
>  %83 = getelementptr i8, i8* %45, i32 %82
>  %84 = bitcast i8* %83 to i32*
>  %85 = load i32, i32* %84
>  %86 = insertelement <8 x i32> undef, i32 %85, i32 0
>  %87 = extractelement <8 x i32> %81, i32 1
>  %88 = getelementptr i8, i8* %45, i32 %87
>  %89 = bitcast i8* %88 to i32*
>  %90 = load i32, i32* %89
>  %91 = insertelement <8 x i32> %86, i32 %90, i32 1
>  %92 = extractelement <8 x i32> %81, i32 2
>  %93 = getelementptr i8, i8* %45, i32 %92
>  %94 = bitcast i8* %93 to i32*
>  %95 = load i32, i32* %94
>  %96 = insertelement <8 x i32> %91, i32 %95, i32 2
>  %97 = extractelement <8 x i32> %81, i32 3
>  %98 = getelementptr i8, i8* %45, i32 %97
>  %99 = bitcast i8* %98 to i32*
>  %100 = load i32, i32* %99
>  %101 = insertelement <8 x i32> %96, i32 %100, i32 3
>  %102 = extractelement <8 x i32> %81, i32 4
>  %103 = getelementptr i8, i8* %45, i32 %102
>  %104 = bitcast i8* %103 to i32*
>  %105 = load i32, i32* %104
>  %106 = insertelement <8 x i32> %101, i32 %105, i32 4
>  %107 = extractelement <8 x i32> %81, i32 5
>  %108 = getelementptr i8, i8* %45, i32 %107
>  %109 = bitcast i8* %108 to i32*
>  %110 = load i32, i32* %109
>  %111 = insertelement <8 x i32> %106, i32 %110, i32 5
>  %112 = extractelement <8 x i32> %81, i32 6
>  %113 = getelementptr i8, i8* %45, i32 %112
>  %114 = bitcast i8* %113 to i32*
>  %115 = load i32, i32* %114
>  %116 = insertelement <8 x i32> %111, i32 %115, i32 6
>  %117 = extractelement <8 x i32> %81, i32 7
>  %118 = getelementptr i8, i8* %45, i32 %117
>  %119 = bitcast i8* %118 to i32*
>  %120 = load i32, i32* %119
>  %121 = insertelement <8 x i32> %116, i32 %120, i32 7
>  %122 = and <8 x i32> %121, <i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215>
>  %123 = sitofp <8 x i32> %122 to <8 x float>
>  %124 = fmul <8 x float> %123, <float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000>
>  store <8 x float> %124, <8 x float>* %sampler0_texel_x_var
>  store <8 x float> %124, <8 x float>* %sampler0_texel_y_var
>  store <8 x float> %124, <8 x float>* %sampler0_texel_z_var
>  store <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, <8 x float>* %sampler0_texel_w_var
>  %sampler0_texel_x = load <8 x float>, <8 x float>* %sampler0_texel_x_var
>  %sampler0_texel_y = load <8 x float>, <8 x float>* %sampler0_texel_y_var
>  %sampler0_texel_z = load <8 x float>, <8 x float>* %sampler0_texel_z_var
>  %sampler0_texel_w = load <8 x float>, <8 x float>* %sampler0_texel_w_var
>  store <8 x float> %sampler0_texel_z, <8 x float>* %output16
>  %125 = load <8 x float>, <8 x float>* %output
>  %126 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 0
>  store <8 x float> %125, <8 x float>* %126
>  %127 = load <8 x float>, <8 x float>* %output11
>  %128 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 1
>  store <8 x float> %127, <8 x float>* %128
>  %129 = load <8 x float>, <8 x float>* %output12
>  %130 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 2
>  store <8 x float> %129, <8 x float>* %130
>  %131 = load <8 x float>, <8 x float>* %output13
>  %132 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 3
>  store <8 x float> %131, <8 x float>* %132
>  %133 = load <8 x float>, <8 x float>* %output16
>  %134 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 2
>  store <8 x float> %133, <8 x float>* %134
>  ret void
>}
>
>FS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		pushq	%rbx
>     5:		andq	$-32, %rsp
>     9:		subq	$32, %rsp
>    13:		movq	768(%rsi), %rax
>    20:		vmovaps	416(%rsi), %ymm1
>    28:		vmovaps	544(%rsi), %ymm2
>    36:		vbroadcastss	(%rax), %ymm3
>    41:		vbroadcastss	16(%rax), %ymm4
>    47:		vbroadcastss	32(%rax), %ymm5
>    53:		movabsq	$140164720619520, %rcx
>    63:		vbroadcastss	(%rcx), %ymm9
>    68:		vsubps	%ymm1, %ymm9, %ymm6
>    72:		vsubps	%ymm2, %ymm6, %ymm6
>    76:		vmulps	%ymm5, %ymm6, %ymm5
>    80:		vmulps	%ymm3, %ymm1, %ymm3
>    84:		vmulps	%ymm4, %ymm2, %ymm4
>    88:		vaddps	%ymm4, %ymm3, %ymm3
>    92:		vaddps	%ymm5, %ymm3, %ymm3
>    96:		vbroadcastss	4(%rax), %ymm4
>   102:		vbroadcastss	20(%rax), %ymm5
>   108:		vbroadcastss	36(%rax), %ymm7
>   114:		vmulps	%ymm7, %ymm6, %ymm6
>   118:		vmulps	%ymm4, %ymm1, %ymm1
>   122:		vmulps	%ymm5, %ymm2, %ymm2
>   126:		vaddps	%ymm2, %ymm1, %ymm1
>   130:		vaddps	%ymm6, %ymm1, %ymm2
>   134:		vmovq	8832(%rdi), %xmm1
>   142:		movslq	8844(%rdi), %rax
>   149:		vmovd	%eax, %xmm4
>   153:		vpsrld	%xmm4, %xmm1, %xmm1
>   157:		movabsq	$140164720619524, %rcx
>   167:		movl	(%rcx), %ecx
>   169:		vmovd	%ecx, %xmm4
>   173:		vpbroadcastd	%xmm4, %xmm5
>   178:		vpmaxsd	%xmm5, %xmm1, %xmm5
>   183:		vpbroadcastd	8864(%rdi,%rax,4), %ymm1
>   193:		movslq	8992(%rdi,%rax,4), %rax
>   201:		addq	8856(%rdi), %rax
>   208:		vpbroadcastd	%xmm5, %ymm6
>   213:		vpbroadcastd	%xmm4, %ymm4
>   218:		vpermd	%ymm5, %ymm4, %ymm7
>   223:		vcvtdq2ps	%xmm5, %xmm5
>   227:		vbroadcastss	%xmm5, %ymm8
>   232:		vpermps	%ymm5, %ymm4, %ymm4
>   237:		vpcmpeqd	%ymm5, %ymm5, %ymm5
>   241:		vpaddd	%ymm5, %ymm6, %ymm6
>   245:		vmulps	%ymm8, %ymm3, %ymm3
>   250:		vcvttps2dq	%ymm3, %ymm3
>   254:		vpminsd	%xmm6, %xmm3, %xmm0
>   259:		vextracti128	$1, %ymm3, %xmm3
>   265:		vextracti128	$1, %ymm6, %xmm6
>   271:		vpminsd	%xmm6, %xmm3, %xmm3
>   276:		vpxor	%xmm6, %xmm6, %xmm6
>   280:		vpmaxsd	%xmm6, %xmm0, %xmm0
>   285:		vpmaxsd	%xmm6, %xmm3, %xmm3
>   290:		vinserti128	$1, %xmm3, %ymm0, %ymm0
>   296:		vpaddd	%ymm5, %ymm7, %ymm3
>   300:		vmulps	%ymm4, %ymm2, %ymm2
>   304:		vcvttps2dq	%ymm2, %ymm2
>   308:		vpminsd	%xmm3, %xmm2, %xmm4
>   313:		vextracti128	$1, %ymm2, %xmm2
>   319:		vextracti128	$1, %ymm3, %xmm3
>   325:		vpminsd	%xmm3, %xmm2, %xmm2
>   330:		vpmaxsd	%xmm6, %xmm4, %xmm3
>   335:		vpmaxsd	%xmm6, %xmm2, %xmm2
>   340:		vinserti128	$1, %xmm2, %ymm3, %ymm2
>   346:		vpslld	$2, %ymm0, %ymm0
>   351:		vpmulld	%ymm1, %ymm2, %ymm1
>   356:		vpaddd	%ymm0, %ymm1, %ymm0
>   360:		vpextrq	$1, %xmm0, %r10
>   366:		vmovq	%xmm0, %rdx
>   371:		movslq	%edx, %r8
>   374:		sarq	$32, %rdx
>   378:		movslq	%r10d, %r9
>   381:		sarq	$32, %r10
>   385:		vextracti128	$1, %ymm0, %xmm0
>   391:		vpextrq	$1, %xmm0, %rdi
>   397:		vmovq	%xmm0, %rcx
>   402:		movslq	%ecx, %r11
>   405:		sarq	$32, %rcx
>   409:		movslq	%edi, %rbx
>   412:		sarq	$32, %rdi
>   416:		vmovd	(%r11,%rax), %xmm0
>   422:		vpinsrd	$1, (%rcx,%rax), %xmm0, %xmm0
>   429:		vpinsrd	$2, (%rbx,%rax), %xmm0, %xmm0
>   436:		vpinsrd	$3, (%rdi,%rax), %xmm0, %xmm0
>   443:		vmovd	(%r8,%rax), %xmm1
>   449:		vpinsrd	$1, (%rdx,%rax), %xmm1, %xmm1
>   456:		vpinsrd	$2, (%r9,%rax), %xmm1, %xmm1
>   463:		vpinsrd	$3, (%r10,%rax), %xmm1, %xmm1
>   470:		vinserti128	$1, %xmm0, %ymm1, %ymm0
>   476:		movabsq	$140164720619528, %rax
>   486:		vpbroadcastd	(%rax), %ymm1
>   491:		vpand	%ymm1, %ymm0, %ymm0
>   495:		vcvtdq2ps	%ymm0, %ymm0
>   499:		movabsq	$140164720619532, %rax
>   509:		vbroadcastss	(%rax), %ymm1
>   514:		vmulps	%ymm1, %ymm0, %ymm0
>   518:		vxorps	%ymm1, %ymm1, %ymm1
>   522:		vmovaps	%ymm1, 832(%rsi)
>   530:		vmovaps	%ymm1, 864(%rsi)
>   538:		vmovaps	%ymm1, 896(%rsi)
>   546:		vmovaps	%ymm9, 928(%rsi)
>   554:		vmovaps	%ymm0, 256(%rsi)
>   562:		leaq	-8(%rbp), %rsp
>   566:		popq	%rbx
>   567:		popq	%rbp
>   568:		vzeroupper
>   571:		retq
>
>
>frag shader  0x7f7aa4620000
>fetch shader 0x7f7aa461e000
>FRAG
>DCL IN[0], GENERIC[9], PERSPECTIVE
>DCL OUT[0], COLOR
>DCL OUT[1], COLOR[1]
>DCL SAMP[0]
>DCL SVIEW[0], 2D, FLOAT
>DCL CONST[0]
>DCL TEMP[0..1], LOCAL
>IMM[0] FLT32 {    1.0000,    -1.0000,     0.0000,     0.0000}
>  0: MOV TEMP[0].xy, IN[0].xyyy
>  1: TEX TEMP[0].x, TEMP[0], SAMP[0], 2D
>  2: FSEQ TEMP[1].x, TEMP[0].xxxx, CONST[0].xxxx
>  3: AND TEMP[1].x, TEMP[1].xxxx, IMM[0].xxxx
>  4: KILL_IF -TEMP[1].xxxx
>  5: MOV TEMP[1].xzw, IMM[0].yyzz
>  6: MOV TEMP[1].y, TEMP[0].xxxx
>  7: MOV OUT[0], TEMP[1]
>  8: MOV OUT[1], TEMP[1]
>  9: END
>TEX[0] = IN[0].x IN[0].y _ _, RES[0], SAMP[0], 2D
>OUT[0].x = -1.000000
>OUT[0].z = 0.000000
>OUT[0].w = 0.000000
>OUT[1].x = -1.000000
>OUT[1].z = 0.000000
>OUT[1].w = 0.000000
>VERT
>PROPERTY NEXT_SHADER 0
>DCL IN[0]
>DCL IN[1]
>DCL OUT[0], POSITION
>DCL OUT[1], GENERIC[9]
>DCL TEMP[0], LOCAL
>  0: MOV TEMP[0].xy, IN[1].xyxx
>  1: MOV OUT[0], IN[0]
>  2: MOV OUT[1], TEMP[0]
>  3: END
>OUT[0].x = IN[0].x
>OUT[0].y = IN[0].y
>OUT[0].z = IN[0].z
>OUT[0].w = IN[0].w
>OUT[1].x = IN[1].x
>OUT[1].y = IN[1].y
>
>; Function Attrs: alignstack(32)
>define void @VS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx) #0 {
>entry:
>  %temp10 = alloca <8 x float>
>  %temp9 = alloca <8 x float>
>  %temp8 = alloca <8 x float>
>  %temp = alloca <8 x float>
>  %output7 = alloca <8 x float>
>  %output6 = alloca <8 x float>
>  %output5 = alloca <8 x float>
>  %output4 = alloca <8 x float>
>  %output3 = alloca <8 x float>
>  %output2 = alloca <8 x float>
>  %output1 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 0
>  %num_vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 1
>  %0 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 0
>  %1 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %0
>  %2 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 0
>  %3 = load <8 x float>, <8 x float>* %2
>  %4 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 1
>  %5 = load <8 x float>, <8 x float>* %4
>  %6 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 2
>  %7 = load <8 x float>, <8 x float>* %6
>  %8 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 3
>  %9 = load <8 x float>, <8 x float>* %8
>  %10 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 0
>  %11 = load <8 x float>, <8 x float>* %10
>  %12 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 1
>  %13 = load <8 x float>, <8 x float>* %12
>  %14 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 2
>  %15 = load i32, i32* %14
>  %16 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 3
>  %17 = load <8 x i32>, <8 x i32>* %16
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output1
>  store <8 x float> zeroinitializer, <8 x float>* %output2
>  store <8 x float> zeroinitializer, <8 x float>* %output3
>  store <8 x float> zeroinitializer, <8 x float>* %output4
>  store <8 x float> zeroinitializer, <8 x float>* %output5
>  store <8 x float> zeroinitializer, <8 x float>* %output6
>  store <8 x float> zeroinitializer, <8 x float>* %output7
>  store <8 x float> zeroinitializer, <8 x float>* %temp
>  store <8 x float> zeroinitializer, <8 x float>* %temp8
>  store <8 x float> zeroinitializer, <8 x float>* %temp9
>  store <8 x float> zeroinitializer, <8 x float>* %temp10
>  store <8 x float> %11, <8 x float>* %temp
>  store <8 x float> %13, <8 x float>* %temp8
>  store <8 x float> %3, <8 x float>* %output
>  store <8 x float> %5, <8 x float>* %output1
>  store <8 x float> %7, <8 x float>* %output2
>  store <8 x float> %9, <8 x float>* %output3
>  %18 = load <8 x float>, <8 x float>* %temp
>  %19 = load <8 x float>, <8 x float>* %temp8
>  %20 = load <8 x float>, <8 x float>* %temp9
>  %21 = load <8 x float>, <8 x float>* %temp10
>  store <8 x float> %18, <8 x float>* %output4
>  store <8 x float> %19, <8 x float>* %output5
>  store <8 x float> %20, <8 x float>* %output6
>  store <8 x float> %21, <8 x float>* %output7
>  %22 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 1
>  %23 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %22
>  %24 = load <8 x float>, <8 x float>* %output
>  %25 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 0
>  store <8 x float> %24, <8 x float>* %25
>  %26 = load <8 x float>, <8 x float>* %output4
>  %27 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 0
>  store <8 x float> %26, <8 x float>* %27
>  %28 = load <8 x float>, <8 x float>* %output1
>  %29 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 1
>  store <8 x float> %28, <8 x float>* %29
>  %30 = load <8 x float>, <8 x float>* %output5
>  %31 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 1
>  store <8 x float> %30, <8 x float>* %31
>  %32 = load <8 x float>, <8 x float>* %output2
>  %33 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 2
>  store <8 x float> %32, <8 x float>* %33
>  %34 = load <8 x float>, <8 x float>* %output6
>  %35 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 2
>  store <8 x float> %34, <8 x float>* %35
>  %36 = load <8 x float>, <8 x float>* %output3
>  %37 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 0, i32 3
>  store <8 x float> %36, <8 x float>* %37
>  %38 = load <8 x float>, <8 x float>* %output7
>  %39 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %23, i32 0, i32 0, i32 1, i32 3
>  store <8 x float> %38, <8 x float>* %39
>  ret void
>}
>
>VS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		andq	$-32, %rsp
>     8:		subq	$32, %rsp
>    12:		movq	(%rsi), %rax
>    15:		movq	8(%rsi), %rcx
>    19:		vmovaps	(%rax), %ymm0
>    23:		vmovaps	32(%rax), %ymm1
>    28:		vmovaps	64(%rax), %ymm2
>    33:		vmovaps	96(%rax), %ymm3
>    38:		vmovaps	128(%rax), %ymm4
>    46:		vmovaps	160(%rax), %ymm5
>    54:		vmovaps	%ymm0, (%rcx)
>    58:		vmovaps	%ymm4, 128(%rcx)
>    66:		vmovaps	%ymm1, 32(%rcx)
>    71:		vmovaps	%ymm5, 160(%rcx)
>    79:		vmovaps	%ymm2, 64(%rcx)
>    84:		vxorps	%ymm0, %ymm0, %ymm0
>    88:		vmovaps	%ymm0, 192(%rcx)
>    96:		vmovaps	%ymm3, 96(%rcx)
>   101:		vmovaps	%ymm0, 224(%rcx)
>   109:		movq	%rbp, %rsp
>   112:		popq	%rbp
>   113:		vzeroupper
>   116:		retq
>
>
>vert shader  0x7f7aa461c000
>
>; Function Attrs: alignstack(32)
>define void @FS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx) #0 {
>entry:
>  %sampler0_texel_w_var = alloca <8 x float>
>  %sampler0_texel_z_var = alloca <8 x float>
>  %sampler0_texel_y_var = alloca <8 x float>
>  %sampler0_texel_x_var = alloca <8 x float>
>  %temp24 = alloca <8 x float>
>  %temp23 = alloca <8 x float>
>  %temp22 = alloca <8 x float>
>  %temp21 = alloca <8 x float>
>  %temp20 = alloca <8 x float>
>  %temp19 = alloca <8 x float>
>  %temp18 = alloca <8 x float>
>  %temp = alloca <8 x float>
>  %output17 = alloca <8 x float>
>  %output16 = alloca <8 x float>
>  %output15 = alloca <8 x float>
>  %output14 = alloca <8 x float>
>  %output13 = alloca <8 x float>
>  %output12 = alloca <8 x float>
>  %output11 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %execution_mask = alloca <8 x i32>
>  %fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 2
>  %num_fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 3
>  %0 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 9
>  %pRawAttribs = load float*, float** %0
>  %1 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 10
>  %pPerspAttribs = load float*, float** %1
>  %2 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i = load <8 x float>, <8 x float>* %2
>  %3 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j = load <8 x float>, <8 x float>* %3
>  %4 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 8, i32 1
>  %5 = load <8 x float>, <8 x float>* %4
>  %w = fdiv <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %5
>  %6 = getelementptr float, float* %pPerspAttribs, i32 0
>  %7 = load float, float* %6
>  %.splatinsert = insertelement <8 x float> undef, float %7, i32 0
>  %.splat = shufflevector <8 x float> %.splatinsert, <8 x float> undef, <8 x i32> zeroinitializer
>  %8 = getelementptr float, float* %pPerspAttribs, i32 4
>  %9 = load float, float* %8
>  %.splatinsert1 = insertelement <8 x float> undef, float %9, i32 0
>  %.splat2 = shufflevector <8 x float> %.splatinsert1, <8 x float> undef, <8 x i32> zeroinitializer
>  %10 = getelementptr float, float* %pPerspAttribs, i32 8
>  %11 = load float, float* %10
>  %.splatinsert3 = insertelement <8 x float> undef, float %11, i32 0
>  %.splat4 = shufflevector <8 x float> %.splatinsert3, <8 x float> undef, <8 x i32> zeroinitializer
>  %12 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %13 = fsub <8 x float> %12, %j
>  %14 = fmul <8 x float> %13, %.splat4
>  %15 = fmul <8 x float> %.splat, %i
>  %16 = fmul <8 x float> %.splat2, %j
>  %17 = fadd <8 x float> %15, %16
>  %18 = fadd <8 x float> %17, %14
>  %19 = fmul <8 x float> %18, %w
>  %20 = getelementptr float, float* %pPerspAttribs, i32 1
>  %21 = load float, float* %20
>  %.splatinsert5 = insertelement <8 x float> undef, float %21, i32 0
>  %.splat6 = shufflevector <8 x float> %.splatinsert5, <8 x float> undef, <8 x i32> zeroinitializer
>  %22 = getelementptr float, float* %pPerspAttribs, i32 5
>  %23 = load float, float* %22
>  %.splatinsert7 = insertelement <8 x float> undef, float %23, i32 0
>  %.splat8 = shufflevector <8 x float> %.splatinsert7, <8 x float> undef, <8 x i32> zeroinitializer
>  %24 = getelementptr float, float* %pPerspAttribs, i32 9
>  %25 = load float, float* %24
>  %.splatinsert9 = insertelement <8 x float> undef, float %25, i32 0
>  %.splat10 = shufflevector <8 x float> %.splatinsert9, <8 x float> undef, <8 x i32> zeroinitializer
>  %26 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i
>  %27 = fsub <8 x float> %26, %j
>  %28 = fmul <8 x float> %27, %.splat10
>  %29 = fmul <8 x float> %.splat6, %i
>  %30 = fmul <8 x float> %.splat8, %j
>  %31 = fadd <8 x float> %29, %30
>  %32 = fadd <8 x float> %31, %28
>  %33 = fmul <8 x float> %32, %w
>  %34 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 3
>  %activeMask = load <8 x i32>, <8 x i32>* %34
>  store <8 x i32> zeroinitializer, <8 x i32>* %execution_mask
>  store <8 x i32> %activeMask, <8 x i32>* %execution_mask
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output11
>  store <8 x float> zeroinitializer, <8 x float>* %output12
>  store <8 x float> zeroinitializer, <8 x float>* %output13
>  store <8 x float> zeroinitializer, <8 x float>* %output14
>  store <8 x float> zeroinitializer, <8 x float>* %output15
>  store <8 x float> zeroinitializer, <8 x float>* %output16
>  store <8 x float> zeroinitializer, <8 x float>* %output17
>  %"&fs_constants[]" = getelementptr [32 x float*], [32 x float*]* %fs_constants, i32 0, i32 0
>  %"fs_constants[]" = load float*, float** %"&fs_constants[]"
>  %"&num_fs_constants[]" = getelementptr [32 x i32], [32 x i32]* %num_fs_constants, i32 0, i32 0
>  %"num_fs_constants[]" = load i32, i32* %"&num_fs_constants[]"
>  store <8 x float> zeroinitializer, <8 x float>* %temp
>  store <8 x float> zeroinitializer, <8 x float>* %temp18
>  store <8 x float> zeroinitializer, <8 x float>* %temp19
>  store <8 x float> zeroinitializer, <8 x float>* %temp20
>  store <8 x float> zeroinitializer, <8 x float>* %temp21
>  store <8 x float> zeroinitializer, <8 x float>* %temp22
>  store <8 x float> zeroinitializer, <8 x float>* %temp23
>  store <8 x float> zeroinitializer, <8 x float>* %temp24
>  store <8 x float> %19, <8 x float>* %temp
>  store <8 x float> %33, <8 x float>* %temp18
>  %35 = load <8 x float>, <8 x float>* %temp
>  %36 = load <8 x float>, <8 x float>* %temp18
>  %37 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 0
>  %context.texture0.width = load i32, i32* %37
>  %context.texture0.row_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 6
>  %context.texture0.img_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 7
>  %38 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 5
>  %context.texture0.base_ptr = load i8*, i8** %38
>  %context.texture0.mip_offsets = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 8
>  %39 = insertelement <4 x i32> undef, i32 %context.texture0.width, i32 0
>  %40 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 1
>  %context.texture0.height = load i32, i32* %40
>  %41 = insertelement <4 x i32> %39, i32 %context.texture0.height, i32 1
>  %42 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 3
>  %context.texture0.first_level = load i32, i32* %42
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_x_var
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_y_var
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_z_var
>  store <8 x float> zeroinitializer, <8 x float>* %sampler0_texel_w_var
>  %43 = insertelement <4 x i32> undef, i32 %context.texture0.first_level, i32 0
>  %44 = shufflevector <4 x i32> %43, <4 x i32> undef, <4 x i32> zeroinitializer
>  %minify = lshr <4 x i32> %41, %44
>  %45 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %46 = getelementptr [16 x i32], [16 x i32]* %context.texture0.row_stride, i32 0, i32 %context.texture0.first_level
>  %47 = load i32, i32* %46
>  %48 = insertelement <8 x i32> undef, i32 %47, i32 0
>  %49 = shufflevector <8 x i32> %48, <8 x i32> undef, <8 x i32> zeroinitializer
>  %50 = getelementptr [16 x i32], [16 x i32]* %context.texture0.mip_offsets, i32 0, i32 %context.texture0.first_level
>  %51 = load i32, i32* %50
>  %52 = getelementptr i8, i8* %context.texture0.base_ptr, i32 %51
>  %53 = shufflevector <4 x i32> %45, <4 x i32> undef, <8 x i32> zeroinitializer
>  %54 = shufflevector <4 x i32> %45, <4 x i32> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %55 = sitofp <4 x i32> %45 to <4 x float>
>  %56 = shufflevector <4 x float> %55, <4 x float> undef, <8 x i32> zeroinitializer
>  %57 = shufflevector <4 x float> %55, <4 x float> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %58 = sub <8 x i32> %53, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %59 = call <8 x float> @llvm.floor.v8f32(<8 x float> %35)
>  %60 = fsub <8 x float> %35, %59
>  %61 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %60, <8 x float> <float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000>)
>  %62 = fmul <8 x float> %61, %56
>  %tex.x.wrapped = fptosi <8 x float> %62 to <8 x i32>
>  %63 = sub <8 x i32> %54, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %64 = call <8 x float> @llvm.floor.v8f32(<8 x float> %36)
>  %65 = fsub <8 x float> %36, %64
>  %66 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %65, <8 x float> <float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000, float 0x3FEFFFFFE0000000>)
>  %67 = fmul <8 x float> %66, %57
>  %tex.y.wrapped = fptosi <8 x float> %67 to <8 x i32>
>  %68 = mul <8 x i32> %tex.x.wrapped, <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4>
>  %69 = mul <8 x i32> %tex.y.wrapped, %49
>  %70 = add <8 x i32> %68, %69
>  %71 = extractelement <8 x i32> %70, i32 0
>  %72 = getelementptr i8, i8* %52, i32 %71
>  %73 = bitcast i8* %72 to i32*
>  %74 = load i32, i32* %73
>  %75 = insertelement <8 x i32> undef, i32 %74, i32 0
>  %76 = extractelement <8 x i32> %70, i32 1
>  %77 = getelementptr i8, i8* %52, i32 %76
>  %78 = bitcast i8* %77 to i32*
>  %79 = load i32, i32* %78
>  %80 = insertelement <8 x i32> %75, i32 %79, i32 1
>  %81 = extractelement <8 x i32> %70, i32 2
>  %82 = getelementptr i8, i8* %52, i32 %81
>  %83 = bitcast i8* %82 to i32*
>  %84 = load i32, i32* %83
>  %85 = insertelement <8 x i32> %80, i32 %84, i32 2
>  %86 = extractelement <8 x i32> %70, i32 3
>  %87 = getelementptr i8, i8* %52, i32 %86
>  %88 = bitcast i8* %87 to i32*
>  %89 = load i32, i32* %88
>  %90 = insertelement <8 x i32> %85, i32 %89, i32 3
>  %91 = extractelement <8 x i32> %70, i32 4
>  %92 = getelementptr i8, i8* %52, i32 %91
>  %93 = bitcast i8* %92 to i32*
>  %94 = load i32, i32* %93
>  %95 = insertelement <8 x i32> %90, i32 %94, i32 4
>  %96 = extractelement <8 x i32> %70, i32 5
>  %97 = getelementptr i8, i8* %52, i32 %96
>  %98 = bitcast i8* %97 to i32*
>  %99 = load i32, i32* %98
>  %100 = insertelement <8 x i32> %95, i32 %99, i32 5
>  %101 = extractelement <8 x i32> %70, i32 6
>  %102 = getelementptr i8, i8* %52, i32 %101
>  %103 = bitcast i8* %102 to i32*
>  %104 = load i32, i32* %103
>  %105 = insertelement <8 x i32> %100, i32 %104, i32 6
>  %106 = extractelement <8 x i32> %70, i32 7
>  %107 = getelementptr i8, i8* %52, i32 %106
>  %108 = bitcast i8* %107 to i32*
>  %109 = load i32, i32* %108
>  %110 = insertelement <8 x i32> %105, i32 %109, i32 7
>  %111 = and <8 x i32> %110, <i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215>
>  %112 = sitofp <8 x i32> %111 to <8 x float>
>  %113 = fmul <8 x float> %112, <float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000>
>  store <8 x float> %113, <8 x float>* %sampler0_texel_x_var
>  store <8 x float> %113, <8 x float>* %sampler0_texel_y_var
>  store <8 x float> %113, <8 x float>* %sampler0_texel_z_var
>  store <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, <8 x float>* %sampler0_texel_w_var
>  %sampler0_texel_x = load <8 x float>, <8 x float>* %sampler0_texel_x_var
>  %sampler0_texel_y = load <8 x float>, <8 x float>* %sampler0_texel_y_var
>  %sampler0_texel_z = load <8 x float>, <8 x float>* %sampler0_texel_z_var
>  %sampler0_texel_w = load <8 x float>, <8 x float>* %sampler0_texel_w_var
>  store <8 x float> %sampler0_texel_x, <8 x float>* %temp
>  %114 = load <8 x float>, <8 x float>* %temp
>  %115 = getelementptr float, float* %"fs_constants[]", i32 0
>  %116 = load float, float* %115
>  %117 = insertelement <8 x float> undef, float %116, i32 0
>  %118 = shufflevector <8 x float> %117, <8 x float> undef, <8 x i32> zeroinitializer
>  %119 = fcmp oeq <8 x float> %114, %118
>  %120 = sext <8 x i1> %119 to <8 x i32>
>  %121 = bitcast <8 x i32> %120 to <8 x float>
>  store <8 x float> %121, <8 x float>* %temp21
>  %122 = load <8 x float>, <8 x float>* %temp21
>  %123 = bitcast <8 x float> %122 to <8 x i32>
>  %124 = and <8 x i32> %123, <i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216>
>  %125 = bitcast <8 x i32> %124 to <8 x float>
>  store <8 x float> %125, <8 x float>* %temp21
>  %126 = load <8 x float>, <8 x float>* %temp21
>  %127 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %126
>  %128 = load <8 x float>, <8 x float>* %temp21
>  %129 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %128
>  %130 = load <8 x float>, <8 x float>* %temp21
>  %131 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %130
>  %132 = load <8 x float>, <8 x float>* %temp21
>  %133 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %132
>  %134 = load <8 x float>, <8 x float>* %temp21
>  %135 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %134
>  %136 = fcmp uge <8 x float> %135, zeroinitializer
>  %137 = sext <8 x i1> %136 to <8 x i32>
>  %138 = load <8 x i32>, <8 x i32>* %execution_mask
>  %139 = and <8 x i32> %138, %137
>  store <8 x i32> %139, <8 x i32>* %execution_mask
>  store <8 x float> <float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00>, <8 x float>* %temp21
>  store <8 x float> zeroinitializer, <8 x float>* %temp23
>  store <8 x float> zeroinitializer, <8 x float>* %temp24
>  %140 = load <8 x float>, <8 x float>* %temp
>  store <8 x float> %140, <8 x float>* %temp22
>  %141 = load <8 x float>, <8 x float>* %temp21
>  %142 = load <8 x float>, <8 x float>* %temp22
>  %143 = load <8 x float>, <8 x float>* %temp23
>  %144 = load <8 x float>, <8 x float>* %temp24
>  store <8 x float> %141, <8 x float>* %output
>  store <8 x float> %142, <8 x float>* %output11
>  store <8 x float> %143, <8 x float>* %output12
>  store <8 x float> %144, <8 x float>* %output13
>  %145 = load <8 x float>, <8 x float>* %temp21
>  %146 = load <8 x float>, <8 x float>* %temp22
>  %147 = load <8 x float>, <8 x float>* %temp23
>  %148 = load <8 x float>, <8 x float>* %temp24
>  store <8 x float> %145, <8 x float>* %output14
>  store <8 x float> %146, <8 x float>* %output15
>  store <8 x float> %147, <8 x float>* %output16
>  store <8 x float> %148, <8 x float>* %output17
>  %149 = load <8 x float>, <8 x float>* %output
>  %150 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 0
>  store <8 x float> %149, <8 x float>* %150
>  %151 = load <8 x float>, <8 x float>* %output11
>  %152 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 1
>  store <8 x float> %151, <8 x float>* %152
>  %153 = load <8 x float>, <8 x float>* %output12
>  %154 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 2
>  store <8 x float> %153, <8 x float>* %154
>  %155 = load <8 x float>, <8 x float>* %output13
>  %156 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 3
>  store <8 x float> %155, <8 x float>* %156
>  %157 = load <8 x float>, <8 x float>* %output14
>  %158 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 0
>  store <8 x float> %157, <8 x float>* %158
>  %159 = load <8 x float>, <8 x float>* %output15
>  %160 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 1
>  store <8 x float> %159, <8 x float>* %160
>  %161 = load <8 x float>, <8 x float>* %output16
>  %162 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 2
>  store <8 x float> %161, <8 x float>* %162
>  %163 = load <8 x float>, <8 x float>* %output17
>  %164 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 3
>  store <8 x float> %163, <8 x float>* %164
>  br label %skip
>
>skip:                                             ; preds = %entry
>  %165 = load <8 x i32>, <8 x i32>* %execution_mask
>  %166 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 3
>  store <8 x i32> %165, <8 x i32>* %166
>  ret void
>}
>
>FS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		pushq	%r14
>     6:		pushq	%rbx
>     7:		andq	$-32, %rsp
>    11:		movq	776(%rsi), %rax
>    18:		vmovaps	288(%rsi), %ymm0
>    26:		vmovaps	416(%rsi), %ymm1
>    34:		movabsq	$140164720594944, %rcx
>    44:		vbroadcastss	(%rcx), %ymm2
>    49:		vdivps	672(%rsi), %ymm2, %ymm3
>    57:		vmovaps	544(%rsi), %ymm4
>    65:		vbroadcastss	(%rax), %ymm5
>    70:		vbroadcastss	16(%rax), %ymm6
>    76:		vbroadcastss	32(%rax), %ymm7
>    82:		vsubps	%ymm1, %ymm2, %ymm2
>    86:		vsubps	%ymm4, %ymm2, %ymm2
>    90:		vmulps	%ymm7, %ymm2, %ymm7
>    94:		vmulps	%ymm5, %ymm1, %ymm5
>    98:		vmulps	%ymm6, %ymm4, %ymm6
>   102:		vaddps	%ymm6, %ymm5, %ymm5
>   106:		vaddps	%ymm7, %ymm5, %ymm5
>   110:		vmulps	%ymm5, %ymm3, %ymm5
>   114:		vbroadcastss	4(%rax), %ymm6
>   120:		vbroadcastss	20(%rax), %ymm7
>   126:		vbroadcastss	36(%rax), %ymm8
>   132:		vmulps	%ymm8, %ymm2, %ymm2
>   137:		vmulps	%ymm6, %ymm1, %ymm1
>   141:		vmulps	%ymm7, %ymm4, %ymm4
>   145:		vaddps	%ymm4, %ymm1, %ymm1
>   149:		vaddps	%ymm2, %ymm1, %ymm1
>   153:		vmulps	%ymm1, %ymm3, %ymm1
>   157:		movq	384(%rdi), %r8
>   164:		vmovq	8832(%rdi), %xmm2
>   172:		movslq	8844(%rdi), %rcx
>   179:		vmovd	%ecx, %xmm3
>   183:		vpsrld	%xmm3, %xmm2, %xmm2
>   187:		movabsq	$140164720594948, %rdx
>   197:		movl	(%rdx), %edx
>   199:		vmovd	%edx, %xmm3
>   203:		vpbroadcastd	%xmm3, %xmm4
>   208:		vpmaxsd	%xmm4, %xmm2, %xmm2
>   213:		vpbroadcastd	8864(%rdi,%rcx,4), %ymm4
>   223:		movslq	8992(%rdi,%rcx,4), %rcx
>   231:		addq	8856(%rdi), %rcx
>   238:		vcvtdq2ps	%xmm2, %xmm2
>   242:		vbroadcastss	%xmm2, %ymm6
>   247:		vpbroadcastd	%xmm3, %ymm3
>   252:		vpermps	%ymm2, %ymm3, %ymm2
>   257:		vroundps	$1, %ymm5, %ymm3
>   263:		vsubps	%ymm3, %ymm5, %ymm3
>   267:		movabsq	$140164720594952, %rdx
>   277:		vbroadcastss	(%rdx), %ymm5
>   282:		vminps	%ymm5, %ymm3, %ymm3
>   286:		vmulps	%ymm3, %ymm6, %ymm3
>   290:		vcvttps2dq	%ymm3, %ymm3
>   294:		vroundps	$1, %ymm1, %ymm6
>   300:		vsubps	%ymm6, %ymm1, %ymm1
>   304:		vminps	%ymm5, %ymm1, %ymm1
>   308:		vmulps	%ymm1, %ymm2, %ymm1
>   312:		vcvttps2dq	%ymm1, %ymm1
>   316:		vpslld	$2, %ymm3, %ymm2
>   321:		vpmulld	%ymm4, %ymm1, %ymm1
>   326:		vpaddd	%ymm2, %ymm1, %ymm1
>   330:		vpextrq	$1, %xmm1, %r11
>   336:		vmovq	%xmm1, %rdi
>   341:		movslq	%edi, %r9
>   344:		sarq	$32, %rdi
>   348:		movslq	%r11d, %r10
>   351:		sarq	$32, %r11
>   355:		vextracti128	$1, %ymm1, %xmm1
>   361:		vpextrq	$1, %xmm1, %rax
>   367:		vmovq	%xmm1, %rdx
>   372:		movslq	%edx, %r14
>   375:		sarq	$32, %rdx
>   379:		movslq	%eax, %rbx
>   382:		sarq	$32, %rax
>   386:		vmovd	(%r14,%rcx), %xmm1
>   392:		vpinsrd	$1, (%rdx,%rcx), %xmm1, %xmm1
>   399:		vpinsrd	$2, (%rbx,%rcx), %xmm1, %xmm1
>   406:		vpinsrd	$3, (%rax,%rcx), %xmm1, %xmm1
>   413:		vmovd	(%r9,%rcx), %xmm2
>   419:		vpinsrd	$1, (%rdi,%rcx), %xmm2, %xmm2
>   426:		vpinsrd	$2, (%r10,%rcx), %xmm2, %xmm2
>   433:		vpinsrd	$3, (%r11,%rcx), %xmm2, %xmm2
>   440:		vinserti128	$1, %xmm1, %ymm2, %ymm1
>   446:		movabsq	$140164720594956, %rax
>   456:		vpbroadcastd	(%rax), %ymm2
>   461:		vpand	%ymm2, %ymm1, %ymm1
>   465:		vcvtdq2ps	%ymm1, %ymm1
>   469:		movabsq	$140164720594960, %rax
>   479:		vbroadcastss	(%rax), %ymm2
>   484:		vmulps	%ymm2, %ymm1, %ymm1
>   488:		vbroadcastss	(%r8), %ymm2
>   493:		vcmpeqps	%ymm2, %ymm1, %ymm2
>   498:		movabsq	$140164720594964, %rax
>   508:		vbroadcastss	(%rax), %ymm3
>   513:		movabsq	$140164720594968, %rax
>   523:		vbroadcastss	(%rax), %ymm4
>   528:		vblendvps	%ymm2, %ymm3, %ymm4, %ymm2
>   534:		vxorps	%ymm4, %ymm4, %ymm4
>   538:		vcmpnltps	%ymm4, %ymm2, %ymm2
>   543:		vandps	%ymm0, %ymm2, %ymm0
>   547:		vmovaps	%ymm3, 832(%rsi)
>   555:		vmovaps	%ymm1, 864(%rsi)
>   563:		vmovaps	%ymm4, 896(%rsi)
>   571:		vmovaps	%ymm4, 928(%rsi)
>   579:		vmovaps	%ymm3, 960(%rsi)
>   587:		vmovaps	%ymm1, 992(%rsi)
>   595:		vmovaps	%ymm4, 1024(%rsi)
>   603:		vmovaps	%ymm4, 1056(%rsi)
>   611:		vmovaps	%ymm0, 288(%rsi)
>   619:		leaq	-16(%rbp), %rsp
>   623:		popq	%rbx
>   624:		popq	%r14
>   626:		popq	%rbp
>   627:		vzeroupper
>   630:		retq
>
>
>frag shader  0x7f7aa461a000
>fetch shader 0x7f7aa4618000
>FRAG
>DCL IN[0], POSITION, LINEAR
>DCL IN[1], FACE, CONSTANT
>DCL IN[2], GENERIC[9], PERSPECTIVE
>DCL IN[3], GENERIC[10], PERSPECTIVE
>DCL OUT[0], COLOR
>DCL OUT[1], COLOR[1]
>DCL SAMP[0]
>DCL SAMP[1]
>DCL SVIEW[0], 2D, FLOAT
>DCL SVIEW[1], 2D, FLOAT
>DCL CONST[7]
>DCL CONST[0..4]
>DCL TEMP[0..1]
>DCL TEMP[2..6], LOCAL
>IMM[0] FLT32 {    0.0000,    -1.0000,    10.0000,     1.0000}
>IMM[1] UINT32 {0, 0, 0, 0}
>IMM[2] INT32 {0, 0, 0, 0}
>  0: MOV TEMP[0], IN[0]
>  1: MAD TEMP[0].y, IN[0], CONST[7].xxxx, CONST[7].yyyy
>  2: FSGE TEMP[1], IN[1], IMM[0].xxxx
>  3: MOV TEMP[2].xy, IN[2].zwzz
>  4: MOV TEMP[2].z, IN[3].xxxx
>  5: F2I TEMP[3].xy, TEMP[0].xyyy
>  6: MOV TEMP[3].xy, TEMP[3].xyyy
>  7: MOV TEMP[3].w, IMM[2].xxxx
>  8: TXF TEMP[3].y, TEMP[3], SAMP[1], 2D
>  9: FSNE TEMP[4].x, TEMP[3].yyyy, IMM[0].yyyy
> 10: FSLT TEMP[5].x, TEMP[3].yyyy, TEMP[0].zzzz
> 11: AND TEMP[4].x, TEMP[4].xxxx, TEMP[5].xxxx
> 12: UIF TEMP[4].xxxx :3
> 13:   MOV TEMP[4].x, IMM[0].yyyy
> 14:   MOV TEMP[4].y, TEMP[3].yyyy
> 15:   MOV TEMP[3].xy, TEMP[4].xyxx
> 16:   MOV OUT[1], TEMP[3]
> 17:   RET
> 18: ELSE :3
> 19:   FSLT TEMP[4].x, TEMP[0].zzzz, IMM[0].zzzz
> 20:   UIF TEMP[4].xxxx :3
> 21:     MOV TEMP[4].x, -TEMP[0].zzzz
> 22:     MOV TEMP[4].y, TEMP[0].zzzz
> 23:     MOV TEMP[3].xy, TEMP[4].xyxx
> 24:     MOV OUT[1], TEMP[3]
> 25:     RET
> 26:   ENDIF
> 27: ENDIF
> 28: DP3 TEMP[4].x, TEMP[2].xyzz, TEMP[2].xyzz
> 29: RSQ TEMP[4].x, TEMP[4].xxxx
> 30: MUL TEMP[4].xyz, TEMP[2].xyzz, TEMP[4].xxxx
> 31: MOV TEMP[5].z, TEMP[4].xyzx
> 32: USEQ TEMP[6].x, TEMP[1].xxxx, IMM[1].xxxx
> 33: UIF TEMP[6].xxxx :3
> 34:   MOV TEMP[5].z, -TEMP[4].xyzx
> 35: ENDIF
> 36: MAX TEMP[4].x, IMM[0].xxxx, TEMP[5].zzzz
> 37: POW TEMP[5].x, TEMP[4].xxxx, CONST[4].xxxx
> 38: MAD TEMP[4].xyz, TEMP[4].xxxx, CONST[2].xyzz, CONST[1].xyzz
> 39: MAD TEMP[4].xyz, TEMP[5].xxxx, CONST[3].xyzz, TEMP[4].xyzz
> 40: MOV TEMP[4].w, CONST[0].xxxx
> 41: MOV_SAT TEMP[4], TEMP[4]
> 42: MOV TEMP[5].xy, IN[2].xyyy
> 43: TEX TEMP[5], TEMP[5], SAMP[0], 2D
> 44: MUL TEMP[4], TEMP[4], TEMP[5]
> 45: FSGE TEMP[2].x, IMM[0].xxxx, TEMP[4].wwww
> 46: AND TEMP[2].x, TEMP[2].xxxx, IMM[0].wwww
> 47: KILL_IF -TEMP[2].xxxx
> 48: MOV OUT[0], TEMP[4]
> 49: MOV OUT[1], TEMP[3]
> 50: END
>TEX[0] = IN[2].x IN[2].y _ _, RES[0], SAMP[0], 2D
>VERT
>PROPERTY NEXT_SHADER 0
>DCL IN[0]
>DCL IN[1]
>DCL IN[2]
>DCL OUT[0], POSITION
>DCL OUT[1], GENERIC[9]
>DCL OUT[2], GENERIC[10]
>DCL CONST[0..6]
>DCL TEMP[0..2], LOCAL
>  0: MUL TEMP[0].xyz, CONST[0].xyzz, IN[1].xxxx
>  1: MAD TEMP[0].xyz, CONST[1].xyzz, IN[1].yyyy, TEMP[0].xyzz
>  2: MAD TEMP[0].xyz, CONST[2].xyzz, IN[1].zzzz, TEMP[0].xyzz
>  3: MUL TEMP[1], CONST[3], IN[0].xxxx
>  4: MAD TEMP[1], CONST[4], IN[0].yyyy, TEMP[1]
>  5: MAD TEMP[1], CONST[5], IN[0].zzzz, TEMP[1]
>  6: MAD TEMP[1], CONST[6], IN[0].wwww, TEMP[1]
>  7: MOV TEMP[2].zw, TEMP[0].yyxy
>  8: MOV TEMP[0].x, TEMP[0].zzzz
>  9: MOV TEMP[2].xy, IN[2].xyxx
> 10: MOV OUT[2], TEMP[0]
> 11: MOV OUT[1], TEMP[2]
> 12: MOV OUT[0], TEMP[1]
> 13: END
>OUT[1].x = IN[2].x
>OUT[1].y = IN[2].y
>
>; Function Attrs: alignstack(32)
>define void @VS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx) #0 {
>entry:
>  %temp22 = alloca <8 x float>
>  %temp21 = alloca <8 x float>
>  %temp20 = alloca <8 x float>
>  %temp19 = alloca <8 x float>
>  %temp18 = alloca <8 x float>
>  %temp17 = alloca <8 x float>
>  %temp16 = alloca <8 x float>
>  %temp15 = alloca <8 x float>
>  %temp14 = alloca <8 x float>
>  %temp13 = alloca <8 x float>
>  %temp12 = alloca <8 x float>
>  %temp = alloca <8 x float>
>  %output11 = alloca <8 x float>
>  %output10 = alloca <8 x float>
>  %output9 = alloca <8 x float>
>  %output8 = alloca <8 x float>
>  %output7 = alloca <8 x float>
>  %output6 = alloca <8 x float>
>  %output5 = alloca <8 x float>
>  %output4 = alloca <8 x float>
>  %output3 = alloca <8 x float>
>  %output2 = alloca <8 x float>
>  %output1 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 0
>  %num_vs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 1
>  %0 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 0
>  %1 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %0
>  %2 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 0
>  %3 = load <8 x float>, <8 x float>* %2
>  %4 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 1
>  %5 = load <8 x float>, <8 x float>* %4
>  %6 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 2
>  %7 = load <8 x float>, <8 x float>* %6
>  %8 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 0, i32 3
>  %9 = load <8 x float>, <8 x float>* %8
>  %10 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 0
>  %11 = load <8 x float>, <8 x float>* %10
>  %12 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 1
>  %13 = load <8 x float>, <8 x float>* %12
>  %14 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 1, i32 2
>  %15 = load <8 x float>, <8 x float>* %14
>  %16 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 2, i32 0
>  %17 = load <8 x float>, <8 x float>* %16
>  %18 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %1, i32 0, i32 0, i32 2, i32 1
>  %19 = load <8 x float>, <8 x float>* %18
>  %20 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 2
>  %21 = load i32, i32* %20
>  %22 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 3
>  %23 = load <8 x i32>, <8 x i32>* %22
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output1
>  store <8 x float> zeroinitializer, <8 x float>* %output2
>  store <8 x float> zeroinitializer, <8 x float>* %output3
>  store <8 x float> zeroinitializer, <8 x float>* %output4
>  store <8 x float> zeroinitializer, <8 x float>* %output5
>  store <8 x float> zeroinitializer, <8 x float>* %output6
>  store <8 x float> zeroinitializer, <8 x float>* %output7
>  store <8 x float> zeroinitializer, <8 x float>* %output8
>  store <8 x float> zeroinitializer, <8 x float>* %output9
>  store <8 x float> zeroinitializer, <8 x float>* %output10
>  store <8 x float> zeroinitializer, <8 x float>* %output11
>  %"&vs_constants[]" = getelementptr [32 x float*], [32 x float*]* %vs_constants, i32 0, i32 0
>  %"vs_constants[]" = load float*, float** %"&vs_constants[]"
>  %"&num_vs_constants[]" = getelementptr [32 x i32], [32 x i32]* %num_vs_constants, i32 0, i32 0
>  %"num_vs_constants[]" = load i32, i32* %"&num_vs_constants[]"
>  store <8 x float> zeroinitializer, <8 x float>* %temp
>  store <8 x float> zeroinitializer, <8 x float>* %temp12
>  store <8 x float> zeroinitializer, <8 x float>* %temp13
>  store <8 x float> zeroinitializer, <8 x float>* %temp14
>  store <8 x float> zeroinitializer, <8 x float>* %temp15
>  store <8 x float> zeroinitializer, <8 x float>* %temp16
>  store <8 x float> zeroinitializer, <8 x float>* %temp17
>  store <8 x float> zeroinitializer, <8 x float>* %temp18
>  store <8 x float> zeroinitializer, <8 x float>* %temp19
>  store <8 x float> zeroinitializer, <8 x float>* %temp20
>  store <8 x float> zeroinitializer, <8 x float>* %temp21
>  store <8 x float> zeroinitializer, <8 x float>* %temp22
>  %24 = getelementptr float, float* %"vs_constants[]", i32 0
>  %25 = load float, float* %24
>  %26 = insertelement <8 x float> undef, float %25, i32 0
>  %27 = shufflevector <8 x float> %26, <8 x float> undef, <8 x i32> zeroinitializer
>  %28 = fmul <8 x float> %27, %11
>  %29 = getelementptr float, float* %"vs_constants[]", i32 1
>  %30 = load float, float* %29
>  %31 = insertelement <8 x float> undef, float %30, i32 0
>  %32 = shufflevector <8 x float> %31, <8 x float> undef, <8 x i32> zeroinitializer
>  %33 = fmul <8 x float> %32, %11
>  %34 = getelementptr float, float* %"vs_constants[]", i32 2
>  %35 = load float, float* %34
>  %36 = insertelement <8 x float> undef, float %35, i32 0
>  %37 = shufflevector <8 x float> %36, <8 x float> undef, <8 x i32> zeroinitializer
>  %38 = fmul <8 x float> %37, %11
>  store <8 x float> %28, <8 x float>* %temp
>  store <8 x float> %33, <8 x float>* %temp12
>  store <8 x float> %38, <8 x float>* %temp13
>  %39 = getelementptr float, float* %"vs_constants[]", i32 4
>  %40 = load float, float* %39
>  %41 = insertelement <8 x float> undef, float %40, i32 0
>  %42 = shufflevector <8 x float> %41, <8 x float> undef, <8 x i32> zeroinitializer
>  %43 = load <8 x float>, <8 x float>* %temp
>  %44 = fmul <8 x float> %42, %13
>  %45 = fadd <8 x float> %44, %43
>  %46 = getelementptr float, float* %"vs_constants[]", i32 5
>  %47 = load float, float* %46
>  %48 = insertelement <8 x float> undef, float %47, i32 0
>  %49 = shufflevector <8 x float> %48, <8 x float> undef, <8 x i32> zeroinitializer
>  %50 = load <8 x float>, <8 x float>* %temp12
>  %51 = fmul <8 x float> %49, %13
>  %52 = fadd <8 x float> %51, %50
>  %53 = getelementptr float, float* %"vs_constants[]", i32 6
>  %54 = load float, float* %53
>  %55 = insertelement <8 x float> undef, float %54, i32 0
>  %56 = shufflevector <8 x float> %55, <8 x float> undef, <8 x i32> zeroinitializer
>  %57 = load <8 x float>, <8 x float>* %temp13
>  %58 = fmul <8 x float> %56, %13
>  %59 = fadd <8 x float> %58, %57
>  store <8 x float> %45, <8 x float>* %temp
>  store <8 x float> %52, <8 x float>* %temp12
>  store <8 x float> %59, <8 x float>* %temp13
>  %60 = getelementptr float, float* %"vs_constants[]", i32 8
>  %61 = load float, float* %60
>  %62 = insertelement <8 x float> undef, float %61, i32 0
>  %63 = shufflevector <8 x float> %62, <8 x float> undef, <8 x i32> zeroinitializer
>  %64 = load <8 x float>, <8 x float>* %temp
>  %65 = fmul <8 x float> %63, %15
>  %66 = fadd <8 x float> %65, %64
>  %67 = getelementptr float, float* %"vs_constants[]", i32 9
>  %68 = load float, float* %67
>  %69 = insertelement <8 x float> undef, float %68, i32 0
>  %70 = shufflevector <8 x float> %69, <8 x float> undef, <8 x i32> zeroinitializer
>  %71 = load <8 x float>, <8 x float>* %temp12
>  %72 = fmul <8 x float> %70, %15
>  %73 = fadd <8 x float> %72, %71
>  %74 = getelementptr float, float* %"vs_constants[]", i32 10
>  %75 = load float, float* %74
>  %76 = insertelement <8 x float> undef, float %75, i32 0
>  %77 = shufflevector <8 x float> %76, <8 x float> undef, <8 x i32> zeroinitializer
>  %78 = load <8 x float>, <8 x float>* %temp13
>  %79 = fmul <8 x float> %77, %15
>  %80 = fadd <8 x float> %79, %78
>  store <8 x float> %66, <8 x float>* %temp
>  store <8 x float> %73, <8 x float>* %temp12
>  store <8 x float> %80, <8 x float>* %temp13
>  %81 = getelementptr float, float* %"vs_constants[]", i32 12
>  %82 = load float, float* %81
>  %83 = insertelement <8 x float> undef, float %82, i32 0
>  %84 = shufflevector <8 x float> %83, <8 x float> undef, <8 x i32> zeroinitializer
>  %85 = fmul <8 x float> %84, %3
>  %86 = getelementptr float, float* %"vs_constants[]", i32 13
>  %87 = load float, float* %86
>  %88 = insertelement <8 x float> undef, float %87, i32 0
>  %89 = shufflevector <8 x float> %88, <8 x float> undef, <8 x i32> zeroinitializer
>  %90 = fmul <8 x float> %89, %3
>  %91 = getelementptr float, float* %"vs_constants[]", i32 14
>  %92 = load float, float* %91
>  %93 = insertelement <8 x float> undef, float %92, i32 0
>  %94 = shufflevector <8 x float> %93, <8 x float> undef, <8 x i32> zeroinitializer
>  %95 = fmul <8 x float> %94, %3
>  %96 = getelementptr float, float* %"vs_constants[]", i32 15
>  %97 = load float, float* %96
>  %98 = insertelement <8 x float> undef, float %97, i32 0
>  %99 = shufflevector <8 x float> %98, <8 x float> undef, <8 x i32> zeroinitializer
>  %100 = fmul <8 x float> %99, %3
>  store <8 x float> %85, <8 x float>* %temp15
>  store <8 x float> %90, <8 x float>* %temp16
>  store <8 x float> %95, <8 x float>* %temp17
>  store <8 x float> %100, <8 x float>* %temp18
>  %101 = getelementptr float, float* %"vs_constants[]", i32 16
>  %102 = load float, float* %101
>  %103 = insertelement <8 x float> undef, float %102, i32 0
>  %104 = shufflevector <8 x float> %103, <8 x float> undef, <8 x i32> zeroinitializer
>  %105 = load <8 x float>, <8 x float>* %temp15
>  %106 = fmul <8 x float> %104, %5
>  %107 = fadd <8 x float> %106, %105
>  %108 = getelementptr float, float* %"vs_constants[]", i32 17
>  %109 = load float, float* %108
>  %110 = insertelement <8 x float> undef, float %109, i32 0
>  %111 = shufflevector <8 x float> %110, <8 x float> undef, <8 x i32> zeroinitializer
>  %112 = load <8 x float>, <8 x float>* %temp16
>  %113 = fmul <8 x float> %111, %5
>  %114 = fadd <8 x float> %113, %112
>  %115 = getelementptr float, float* %"vs_constants[]", i32 18
>  %116 = load float, float* %115
>  %117 = insertelement <8 x float> undef, float %116, i32 0
>  %118 = shufflevector <8 x float> %117, <8 x float> undef, <8 x i32> zeroinitializer
>  %119 = load <8 x float>, <8 x float>* %temp17
>  %120 = fmul <8 x float> %118, %5
>  %121 = fadd <8 x float> %120, %119
>  %122 = getelementptr float, float* %"vs_constants[]", i32 19
>  %123 = load float, float* %122
>  %124 = insertelement <8 x float> undef, float %123, i32 0
>  %125 = shufflevector <8 x float> %124, <8 x float> undef, <8 x i32> zeroinitializer
>  %126 = load <8 x float>, <8 x float>* %temp18
>  %127 = fmul <8 x float> %125, %5
>  %128 = fadd <8 x float> %127, %126
>  store <8 x float> %107, <8 x float>* %temp15
>  store <8 x float> %114, <8 x float>* %temp16
>  store <8 x float> %121, <8 x float>* %temp17
>  store <8 x float> %128, <8 x float>* %temp18
>  %129 = getelementptr float, float* %"vs_constants[]", i32 20
>  %130 = load float, float* %129
>  %131 = insertelement <8 x float> undef, float %130, i32 0
>  %132 = shufflevector <8 x float> %131, <8 x float> undef, <8 x i32> zeroinitializer
>  %133 = load <8 x float>, <8 x float>* %temp15
>  %134 = fmul <8 x float> %132, %7
>  %135 = fadd <8 x float> %134, %133
>  %136 = getelementptr float, float* %"vs_constants[]", i32 21
>  %137 = load float, float* %136
>  %138 = insertelement <8 x float> undef, float %137, i32 0
>  %139 = shufflevector <8 x float> %138, <8 x float> undef, <8 x i32> zeroinitializer
>  %140 = load <8 x float>, <8 x float>* %temp16
>  %141 = fmul <8 x float> %139, %7
>  %142 = fadd <8 x float> %141, %140
>  %143 = getelementptr float, float* %"vs_constants[]", i32 22
>  %144 = load float, float* %143
>  %145 = insertelement <8 x float> undef, float %144, i32 0
>  %146 = shufflevector <8 x float> %145, <8 x float> undef, <8 x i32> zeroinitializer
>  %147 = load <8 x float>, <8 x float>* %temp17
>  %148 = fmul <8 x float> %146, %7
>  %149 = fadd <8 x float> %148, %147
>  %150 = getelementptr float, float* %"vs_constants[]", i32 23
>  %151 = load float, float* %150
>  %152 = insertelement <8 x float> undef, float %151, i32 0
>  %153 = shufflevector <8 x float> %152, <8 x float> undef, <8 x i32> zeroinitializer
>  %154 = load <8 x float>, <8 x float>* %temp18
>  %155 = fmul <8 x float> %153, %7
>  %156 = fadd <8 x float> %155, %154
>  store <8 x float> %135, <8 x float>* %temp15
>  store <8 x float> %142, <8 x float>* %temp16
>  store <8 x float> %149, <8 x float>* %temp17
>  store <8 x float> %156, <8 x float>* %temp18
>  %157 = getelementptr float, float* %"vs_constants[]", i32 24
>  %158 = load float, float* %157
>  %159 = insertelement <8 x float> undef, float %158, i32 0
>  %160 = shufflevector <8 x float> %159, <8 x float> undef, <8 x i32> zeroinitializer
>  %161 = load <8 x float>, <8 x float>* %temp15
>  %162 = fmul <8 x float> %160, %9
>  %163 = fadd <8 x float> %162, %161
>  %164 = getelementptr float, float* %"vs_constants[]", i32 25
>  %165 = load float, float* %164
>  %166 = insertelement <8 x float> undef, float %165, i32 0
>  %167 = shufflevector <8 x float> %166, <8 x float> undef, <8 x i32> zeroinitializer
>  %168 = load <8 x float>, <8 x float>* %temp16
>  %169 = fmul <8 x float> %167, %9
>  %170 = fadd <8 x float> %169, %168
>  %171 = getelementptr float, float* %"vs_constants[]", i32 26
>  %172 = load float, float* %171
>  %173 = insertelement <8 x float> undef, float %172, i32 0
>  %174 = shufflevector <8 x float> %173, <8 x float> undef, <8 x i32> zeroinitializer
>  %175 = load <8 x float>, <8 x float>* %temp17
>  %176 = fmul <8 x float> %174, %9
>  %177 = fadd <8 x float> %176, %175
>  %178 = getelementptr float, float* %"vs_constants[]", i32 27
>  %179 = load float, float* %178
>  %180 = insertelement <8 x float> undef, float %179, i32 0
>  %181 = shufflevector <8 x float> %180, <8 x float> undef, <8 x i32> zeroinitializer
>  %182 = load <8 x float>, <8 x float>* %temp18
>  %183 = fmul <8 x float> %181, %9
>  %184 = fadd <8 x float> %183, %182
>  store <8 x float> %163, <8 x float>* %temp15
>  store <8 x float> %170, <8 x float>* %temp16
>  store <8 x float> %177, <8 x float>* %temp17
>  store <8 x float> %184, <8 x float>* %temp18
>  %185 = load <8 x float>, <8 x float>* %temp
>  %186 = load <8 x float>, <8 x float>* %temp12
>  store <8 x float> %185, <8 x float>* %temp21
>  store <8 x float> %186, <8 x float>* %temp22
>  %187 = load <8 x float>, <8 x float>* %temp13
>  store <8 x float> %187, <8 x float>* %temp
>  store <8 x float> %17, <8 x float>* %temp19
>  store <8 x float> %19, <8 x float>* %temp20
>  %188 = load <8 x float>, <8 x float>* %temp
>  %189 = load <8 x float>, <8 x float>* %temp12
>  %190 = load <8 x float>, <8 x float>* %temp13
>  %191 = load <8 x float>, <8 x float>* %temp14
>  store <8 x float> %188, <8 x float>* %output8
>  store <8 x float> %189, <8 x float>* %output9
>  store <8 x float> %190, <8 x float>* %output10
>  store <8 x float> %191, <8 x float>* %output11
>  %192 = load <8 x float>, <8 x float>* %temp19
>  %193 = load <8 x float>, <8 x float>* %temp20
>  %194 = load <8 x float>, <8 x float>* %temp21
>  %195 = load <8 x float>, <8 x float>* %temp22
>  store <8 x float> %192, <8 x float>* %output4
>  store <8 x float> %193, <8 x float>* %output5
>  store <8 x float> %194, <8 x float>* %output6
>  store <8 x float> %195, <8 x float>* %output7
>  %196 = load <8 x float>, <8 x float>* %temp15
>  %197 = load <8 x float>, <8 x float>* %temp16
>  %198 = load <8 x float>, <8 x float>* %temp17
>  %199 = load <8 x float>, <8 x float>* %temp18
>  store <8 x float> %196, <8 x float>* %output
>  store <8 x float> %197, <8 x float>* %output1
>  store <8 x float> %198, <8 x float>* %output2
>  store <8 x float> %199, <8 x float>* %output3
>  %200 = getelementptr { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }, { { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }*, i32, <8 x i32>, <8 x i32> }* %vsCtx, i32 0, i32 1
>  %201 = load { [38 x [4 x <8 x float>]] }*, { [38 x [4 x <8 x float>]] }** %200
>  %202 = load <8 x float>, <8 x float>* %output
>  %203 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 0, i32 0
>  store <8 x float> %202, <8 x float>* %203
>  %204 = load <8 x float>, <8 x float>* %output4
>  %205 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 1, i32 0
>  store <8 x float> %204, <8 x float>* %205
>  %206 = load <8 x float>, <8 x float>* %output8
>  %207 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 2, i32 0
>  store <8 x float> %206, <8 x float>* %207
>  %208 = load <8 x float>, <8 x float>* %output1
>  %209 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 0, i32 1
>  store <8 x float> %208, <8 x float>* %209
>  %210 = load <8 x float>, <8 x float>* %output5
>  %211 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 1, i32 1
>  store <8 x float> %210, <8 x float>* %211
>  %212 = load <8 x float>, <8 x float>* %output9
>  %213 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 2, i32 1
>  store <8 x float> %212, <8 x float>* %213
>  %214 = load <8 x float>, <8 x float>* %output2
>  %215 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 0, i32 2
>  store <8 x float> %214, <8 x float>* %215
>  %216 = load <8 x float>, <8 x float>* %output6
>  %217 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 1, i32 2
>  store <8 x float> %216, <8 x float>* %217
>  %218 = load <8 x float>, <8 x float>* %output10
>  %219 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 2, i32 2
>  store <8 x float> %218, <8 x float>* %219
>  %220 = load <8 x float>, <8 x float>* %output3
>  %221 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 0, i32 3
>  store <8 x float> %220, <8 x float>* %221
>  %222 = load <8 x float>, <8 x float>* %output7
>  %223 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 1, i32 3
>  store <8 x float> %222, <8 x float>* %223
>  %224 = load <8 x float>, <8 x float>* %output11
>  %225 = getelementptr { [38 x [4 x <8 x float>]] }, { [38 x [4 x <8 x float>]] }* %201, i32 0, i32 0, i32 2, i32 3
>  store <8 x float> %224, <8 x float>* %225
>  ret void
>}
>
>VS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		andq	$-32, %rsp
>     8:		subq	$32, %rsp
>    12:		movq	(%rsi), %rcx
>    15:		movq	8(%rsi), %rax
>    19:		vmovaps	(%rcx), %ymm8
>    23:		vmovaps	32(%rcx), %ymm7
>    28:		vmovaps	64(%rcx), %ymm3
>    33:		vmovaps	96(%rcx), %ymm1
>    38:		vmovaps	128(%rcx), %ymm4
>    46:		vmovaps	160(%rcx), %ymm5
>    54:		vmovaps	192(%rcx), %ymm6
>    62:		vmovaps	256(%rcx), %ymm2
>    70:		vmovaps	288(%rcx), %ymm0
>    78:		movq	(%rdi), %rcx
>    81:		vbroadcastss	(%rcx), %ymm9
>    86:		vmulps	%ymm9, %ymm4, %ymm9
>    91:		vbroadcastss	4(%rcx), %ymm10
>    97:		vmulps	%ymm10, %ymm4, %ymm10
>   102:		vbroadcastss	8(%rcx), %ymm11
>   108:		vmulps	%ymm11, %ymm4, %ymm4
>   113:		vbroadcastss	16(%rcx), %ymm11
>   119:		vmulps	%ymm11, %ymm5, %ymm11
>   124:		vaddps	%ymm11, %ymm9, %ymm9
>   129:		vbroadcastss	20(%rcx), %ymm11
>   135:		vmulps	%ymm11, %ymm5, %ymm11
>   140:		vaddps	%ymm11, %ymm10, %ymm10
>   145:		vbroadcastss	24(%rcx), %ymm11
>   151:		vmulps	%ymm11, %ymm5, %ymm5
>   156:		vaddps	%ymm5, %ymm4, %ymm11
>   160:		vbroadcastss	32(%rcx), %ymm4
>   166:		vmulps	%ymm4, %ymm6, %ymm4
>   170:		vaddps	%ymm4, %ymm9, %ymm4
>   174:		vbroadcastss	36(%rcx), %ymm5
>   180:		vmulps	%ymm5, %ymm6, %ymm5
>   184:		vaddps	%ymm5, %ymm10, %ymm5
>   188:		vbroadcastss	40(%rcx), %ymm9
>   194:		vmulps	%ymm9, %ymm6, %ymm6
>   199:		vaddps	%ymm6, %ymm11, %ymm6
>   203:		vbroadcastss	48(%rcx), %ymm9
>   209:		vmulps	%ymm9, %ymm8, %ymm9
>   214:		vbroadcastss	52(%rcx), %ymm10
>   220:		vmulps	%ymm10, %ymm8, %ymm10
>   225:		vbroadcastss	56(%rcx), %ymm11
>   231:		vmulps	%ymm11, %ymm8, %ymm11
>   236:		vbroadcastss	60(%rcx), %ymm12
>   242:		vmulps	%ymm12, %ymm8, %ymm8
>   247:		vbroadcastss	64(%rcx), %ymm12
>   253:		vmulps	%ymm12, %ymm7, %ymm12
>   258:		vaddps	%ymm12, %ymm9, %ymm9
>   263:		vbroadcastss	68(%rcx), %ymm12
>   269:		vmulps	%ymm12, %ymm7, %ymm12
>   274:		vaddps	%ymm12, %ymm10, %ymm10
>   279:		vbroadcastss	72(%rcx), %ymm12
>   285:		vmulps	%ymm12, %ymm7, %ymm12
>   290:		vaddps	%ymm12, %ymm11, %ymm11
>   295:		vbroadcastss	76(%rcx), %ymm12
>   301:		vmulps	%ymm12, %ymm7, %ymm7
>   306:		vaddps	%ymm7, %ymm8, %ymm7
>   310:		vbroadcastss	80(%rcx), %ymm8
>   316:		vmulps	%ymm8, %ymm3, %ymm8
>   321:		vaddps	%ymm8, %ymm9, %ymm8
>   326:		vbroadcastss	84(%rcx), %ymm9
>   332:		vmulps	%ymm9, %ymm3, %ymm9
>   337:		vaddps	%ymm9, %ymm10, %ymm9
>   342:		vbroadcastss	88(%rcx), %ymm10
>   348:		vmulps	%ymm10, %ymm3, %ymm10
>   353:		vaddps	%ymm10, %ymm11, %ymm10
>   358:		vbroadcastss	92(%rcx), %ymm11
>   364:		vmulps	%ymm11, %ymm3, %ymm3
>   369:		vaddps	%ymm3, %ymm7, %ymm3
>   373:		vbroadcastss	96(%rcx), %ymm7
>   379:		vmulps	%ymm7, %ymm1, %ymm7
>   383:		vaddps	%ymm7, %ymm8, %ymm7
>   387:		vbroadcastss	100(%rcx), %ymm8
>   393:		vmulps	%ymm8, %ymm1, %ymm8
>   398:		vaddps	%ymm8, %ymm9, %ymm8
>   403:		vbroadcastss	104(%rcx), %ymm9
>   409:		vmulps	%ymm9, %ymm1, %ymm9
>   414:		vaddps	%ymm9, %ymm10, %ymm9
>   419:		vbroadcastss	108(%rcx), %ymm10
>   425:		vmulps	%ymm10, %ymm1, %ymm1
>   430:		vaddps	%ymm1, %ymm3, %ymm1
>   434:		vmovaps	%ymm7, (%rax)
>   438:		vmovaps	%ymm2, 128(%rax)
>   446:		vmovaps	%ymm6, 256(%rax)
>   454:		vmovaps	%ymm8, 32(%rax)
>   459:		vmovaps	%ymm0, 160(%rax)
>   467:		vmovaps	%ymm5, 288(%rax)
>   475:		vmovaps	%ymm9, 64(%rax)
>   480:		vmovaps	%ymm4, 192(%rax)
>   488:		vmovaps	%ymm6, 320(%rax)
>   496:		vmovaps	%ymm1, 96(%rax)
>   501:		vmovaps	%ymm5, 224(%rax)
>   509:		vxorps	%ymm0, %ymm0, %ymm0
>   513:		vmovaps	%ymm0, 352(%rax)
>   521:		movq	%rbp, %rsp
>   524:		popq	%rbp
>   525:		vzeroupper
>   528:		retq
>
>
>vert shader  0x7f7aa4616000
>
>; Function Attrs: alignstack(32)
>define void @FS({ [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx) #0 {
>entry:
>  %packed_var91 = alloca <16 x i8>
>  %packed_var = alloca <16 x i8>
>  %temp73 = alloca <8 x float>
>  %temp72 = alloca <8 x float>
>  %temp71 = alloca <8 x float>
>  %temp70 = alloca <8 x float>
>  %temp69 = alloca <8 x float>
>  %temp68 = alloca <8 x float>
>  %temp67 = alloca <8 x float>
>  %temp66 = alloca <8 x float>
>  %temp65 = alloca <8 x float>
>  %temp64 = alloca <8 x float>
>  %temp63 = alloca <8 x float>
>  %temp62 = alloca <8 x float>
>  %temp61 = alloca <8 x float>
>  %temp60 = alloca <8 x float>
>  %temp59 = alloca <8 x float>
>  %temp58 = alloca <8 x float>
>  %temp57 = alloca <8 x float>
>  %temp56 = alloca <8 x float>
>  %temp55 = alloca <8 x float>
>  %temp54 = alloca <8 x float>
>  %temp53 = alloca <8 x float>
>  %temp52 = alloca <8 x float>
>  %temp51 = alloca <8 x float>
>  %temp50 = alloca <8 x float>
>  %temp49 = alloca <8 x float>
>  %temp48 = alloca <8 x float>
>  %temp47 = alloca <8 x float>
>  %temp = alloca <8 x float>
>  %output42 = alloca <8 x float>
>  %output41 = alloca <8 x float>
>  %output40 = alloca <8 x float>
>  %output39 = alloca <8 x float>
>  %output38 = alloca <8 x float>
>  %output37 = alloca <8 x float>
>  %output36 = alloca <8 x float>
>  %output = alloca <8 x float>
>  %looplimiter = alloca i32
>  %execution_mask = alloca <8 x i32>
>  %fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 2
>  %num_fs_constants = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 3
>  %0 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 9
>  %pRawAttribs = load float*, float** %0
>  %1 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 10
>  %pPerspAttribs = load float*, float** %1
>  %2 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i = load <8 x float>, <8 x float>* %2
>  %3 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j = load <8 x float>, <8 x float>* %3
>  %4 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 0, i32 1
>  %vX = load <8 x float>, <8 x float>* %4
>  %5 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 1, i32 1
>  %vY = load <8 x float>, <8 x float>* %5
>  %6 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 2
>  %vZ = load <8 x float>, <8 x float>* %6
>  %7 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 8, i32 1
>  %vOneOverW = load <8 x float>, <8 x float>* %7
>  %8 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i1 = load <8 x float>, <8 x float>* %8
>  %9 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j2 = load <8 x float>, <8 x float>* %9
>  %10 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 18
>  %11 = load i32, i32* %10
>  %12 = uitofp i32 %11 to float
>  %13 = fmul float %12, 2.000000e+00
>  %14 = fsub float %13, 1.000000e+00
>  %vFrontFace.splatinsert = insertelement <8 x float> undef, float %14, i32 0
>  %vFrontFace.splat = shufflevector <8 x float> %vFrontFace.splatinsert, <8 x float> undef, <8 x i32> zeroinitializer
>  %15 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i3 = load <8 x float>, <8 x float>* %15
>  %16 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j4 = load <8 x float>, <8 x float>* %16
>  %17 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 8, i32 1
>  %18 = load <8 x float>, <8 x float>* %17
>  %w = fdiv <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %18
>  %19 = getelementptr float, float* %pPerspAttribs, i32 0
>  %20 = load float, float* %19
>  %.splatinsert = insertelement <8 x float> undef, float %20, i32 0
>  %.splat = shufflevector <8 x float> %.splatinsert, <8 x float> undef, <8 x i32> zeroinitializer
>  %21 = getelementptr float, float* %pPerspAttribs, i32 4
>  %22 = load float, float* %21
>  %.splatinsert5 = insertelement <8 x float> undef, float %22, i32 0
>  %.splat6 = shufflevector <8 x float> %.splatinsert5, <8 x float> undef, <8 x i32> zeroinitializer
>  %23 = getelementptr float, float* %pPerspAttribs, i32 8
>  %24 = load float, float* %23
>  %.splatinsert7 = insertelement <8 x float> undef, float %24, i32 0
>  %.splat8 = shufflevector <8 x float> %.splatinsert7, <8 x float> undef, <8 x i32> zeroinitializer
>  %25 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i3
>  %26 = fsub <8 x float> %25, %j4
>  %27 = fmul <8 x float> %26, %.splat8
>  %28 = fmul <8 x float> %.splat, %i3
>  %29 = fmul <8 x float> %.splat6, %j4
>  %30 = fadd <8 x float> %28, %29
>  %31 = fadd <8 x float> %30, %27
>  %32 = fmul <8 x float> %31, %w
>  %33 = getelementptr float, float* %pPerspAttribs, i32 1
>  %34 = load float, float* %33
>  %.splatinsert9 = insertelement <8 x float> undef, float %34, i32 0
>  %.splat10 = shufflevector <8 x float> %.splatinsert9, <8 x float> undef, <8 x i32> zeroinitializer
>  %35 = getelementptr float, float* %pPerspAttribs, i32 5
>  %36 = load float, float* %35
>  %.splatinsert11 = insertelement <8 x float> undef, float %36, i32 0
>  %.splat12 = shufflevector <8 x float> %.splatinsert11, <8 x float> undef, <8 x i32> zeroinitializer
>  %37 = getelementptr float, float* %pPerspAttribs, i32 9
>  %38 = load float, float* %37
>  %.splatinsert13 = insertelement <8 x float> undef, float %38, i32 0
>  %.splat14 = shufflevector <8 x float> %.splatinsert13, <8 x float> undef, <8 x i32> zeroinitializer
>  %39 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i3
>  %40 = fsub <8 x float> %39, %j4
>  %41 = fmul <8 x float> %40, %.splat14
>  %42 = fmul <8 x float> %.splat10, %i3
>  %43 = fmul <8 x float> %.splat12, %j4
>  %44 = fadd <8 x float> %42, %43
>  %45 = fadd <8 x float> %44, %41
>  %46 = fmul <8 x float> %45, %w
>  %47 = getelementptr float, float* %pPerspAttribs, i32 2
>  %48 = load float, float* %47
>  %.splatinsert15 = insertelement <8 x float> undef, float %48, i32 0
>  %.splat16 = shufflevector <8 x float> %.splatinsert15, <8 x float> undef, <8 x i32> zeroinitializer
>  %49 = getelementptr float, float* %pPerspAttribs, i32 6
>  %50 = load float, float* %49
>  %.splatinsert17 = insertelement <8 x float> undef, float %50, i32 0
>  %.splat18 = shufflevector <8 x float> %.splatinsert17, <8 x float> undef, <8 x i32> zeroinitializer
>  %51 = getelementptr float, float* %pPerspAttribs, i32 10
>  %52 = load float, float* %51
>  %.splatinsert19 = insertelement <8 x float> undef, float %52, i32 0
>  %.splat20 = shufflevector <8 x float> %.splatinsert19, <8 x float> undef, <8 x i32> zeroinitializer
>  %53 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i3
>  %54 = fsub <8 x float> %53, %j4
>  %55 = fmul <8 x float> %54, %.splat20
>  %56 = fmul <8 x float> %.splat16, %i3
>  %57 = fmul <8 x float> %.splat18, %j4
>  %58 = fadd <8 x float> %56, %57
>  %59 = fadd <8 x float> %58, %55
>  %60 = fmul <8 x float> %59, %w
>  %61 = getelementptr float, float* %pPerspAttribs, i32 3
>  %62 = load float, float* %61
>  %.splatinsert21 = insertelement <8 x float> undef, float %62, i32 0
>  %.splat22 = shufflevector <8 x float> %.splatinsert21, <8 x float> undef, <8 x i32> zeroinitializer
>  %63 = getelementptr float, float* %pPerspAttribs, i32 7
>  %64 = load float, float* %63
>  %.splatinsert23 = insertelement <8 x float> undef, float %64, i32 0
>  %.splat24 = shufflevector <8 x float> %.splatinsert23, <8 x float> undef, <8 x i32> zeroinitializer
>  %65 = getelementptr float, float* %pPerspAttribs, i32 11
>  %66 = load float, float* %65
>  %.splatinsert25 = insertelement <8 x float> undef, float %66, i32 0
>  %.splat26 = shufflevector <8 x float> %.splatinsert25, <8 x float> undef, <8 x i32> zeroinitializer
>  %67 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i3
>  %68 = fsub <8 x float> %67, %j4
>  %69 = fmul <8 x float> %68, %.splat26
>  %70 = fmul <8 x float> %.splat22, %i3
>  %71 = fmul <8 x float> %.splat24, %j4
>  %72 = fadd <8 x float> %70, %71
>  %73 = fadd <8 x float> %72, %69
>  %74 = fmul <8 x float> %73, %w
>  %75 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 6, i32 1
>  %i27 = load <8 x float>, <8 x float>* %75
>  %76 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 7, i32 1
>  %j28 = load <8 x float>, <8 x float>* %76
>  %77 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 8, i32 1
>  %78 = load <8 x float>, <8 x float>* %77
>  %w29 = fdiv <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %78
>  %79 = getelementptr float, float* %pPerspAttribs, i32 12
>  %80 = load float, float* %79
>  %.splatinsert30 = insertelement <8 x float> undef, float %80, i32 0
>  %.splat31 = shufflevector <8 x float> %.splatinsert30, <8 x float> undef, <8 x i32> zeroinitializer
>  %81 = getelementptr float, float* %pPerspAttribs, i32 16
>  %82 = load float, float* %81
>  %.splatinsert32 = insertelement <8 x float> undef, float %82, i32 0
>  %.splat33 = shufflevector <8 x float> %.splatinsert32, <8 x float> undef, <8 x i32> zeroinitializer
>  %83 = getelementptr float, float* %pPerspAttribs, i32 20
>  %84 = load float, float* %83
>  %.splatinsert34 = insertelement <8 x float> undef, float %84, i32 0
>  %.splat35 = shufflevector <8 x float> %.splatinsert34, <8 x float> undef, <8 x i32> zeroinitializer
>  %85 = fsub <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %i27
>  %86 = fsub <8 x float> %85, %j28
>  %87 = fmul <8 x float> %86, %.splat35
>  %88 = fmul <8 x float> %.splat31, %i27
>  %89 = fmul <8 x float> %.splat33, %j28
>  %90 = fadd <8 x float> %88, %89
>  %91 = fadd <8 x float> %90, %87
>  %92 = fmul <8 x float> %91, %w29
>  %93 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 3
>  %activeMask = load <8 x i32>, <8 x i32>* %93
>  store <8 x i32> zeroinitializer, <8 x i32>* %execution_mask
>  store <8 x i32> %activeMask, <8 x i32>* %execution_mask
>  store i32 0, i32* %looplimiter
>  store i32 65535, i32* %looplimiter
>  store <8 x float> zeroinitializer, <8 x float>* %output
>  store <8 x float> zeroinitializer, <8 x float>* %output36
>  store <8 x float> zeroinitializer, <8 x float>* %output37
>  store <8 x float> zeroinitializer, <8 x float>* %output38
>  store <8 x float> zeroinitializer, <8 x float>* %output39
>  store <8 x float> zeroinitializer, <8 x float>* %output40
>  store <8 x float> zeroinitializer, <8 x float>* %output41
>  store <8 x float> zeroinitializer, <8 x float>* %output42
>  %"&fs_constants[]" = getelementptr [32 x float*], [32 x float*]* %fs_constants, i32 0, i32 0
>  %"fs_constants[]" = load float*, float** %"&fs_constants[]"
>  %"&num_fs_constants[]" = getelementptr [32 x i32], [32 x i32]* %num_fs_constants, i32 0, i32 0
>  %"num_fs_constants[]" = load i32, i32* %"&num_fs_constants[]"
>  %"&fs_constants[]43" = getelementptr [32 x float*], [32 x float*]* %fs_constants, i32 0, i32 0
>  %"fs_constants[]44" = load float*, float** %"&fs_constants[]43"
>  %"&num_fs_constants[]45" = getelementptr [32 x i32], [32 x i32]* %num_fs_constants, i32 0, i32 0
>  %"num_fs_constants[]46" = load i32, i32* %"&num_fs_constants[]45"
>  store <8 x float> zeroinitializer, <8 x float>* %temp
>  store <8 x float> zeroinitializer, <8 x float>* %temp47
>  store <8 x float> zeroinitializer, <8 x float>* %temp48
>  store <8 x float> zeroinitializer, <8 x float>* %temp49
>  store <8 x float> zeroinitializer, <8 x float>* %temp50
>  store <8 x float> zeroinitializer, <8 x float>* %temp51
>  store <8 x float> zeroinitializer, <8 x float>* %temp52
>  store <8 x float> zeroinitializer, <8 x float>* %temp53
>  store <8 x float> zeroinitializer, <8 x float>* %temp54
>  store <8 x float> zeroinitializer, <8 x float>* %temp55
>  store <8 x float> zeroinitializer, <8 x float>* %temp56
>  store <8 x float> zeroinitializer, <8 x float>* %temp57
>  store <8 x float> zeroinitializer, <8 x float>* %temp58
>  store <8 x float> zeroinitializer, <8 x float>* %temp59
>  store <8 x float> zeroinitializer, <8 x float>* %temp60
>  store <8 x float> zeroinitializer, <8 x float>* %temp61
>  store <8 x float> zeroinitializer, <8 x float>* %temp62
>  store <8 x float> zeroinitializer, <8 x float>* %temp63
>  store <8 x float> zeroinitializer, <8 x float>* %temp64
>  store <8 x float> zeroinitializer, <8 x float>* %temp65
>  store <8 x float> zeroinitializer, <8 x float>* %temp66
>  store <8 x float> zeroinitializer, <8 x float>* %temp67
>  store <8 x float> zeroinitializer, <8 x float>* %temp68
>  store <8 x float> zeroinitializer, <8 x float>* %temp69
>  store <8 x float> zeroinitializer, <8 x float>* %temp70
>  store <8 x float> zeroinitializer, <8 x float>* %temp71
>  store <8 x float> zeroinitializer, <8 x float>* %temp72
>  store <8 x float> zeroinitializer, <8 x float>* %temp73
>  store <8 x float> %vX, <8 x float>* %temp
>  store <8 x float> %vY, <8 x float>* %temp47
>  store <8 x float> %vZ, <8 x float>* %temp48
>  store <8 x float> %vOneOverW, <8 x float>* %temp49
>  %94 = getelementptr float, float* %"fs_constants[]44", i32 28
>  %95 = load float, float* %94
>  %96 = insertelement <8 x float> undef, float %95, i32 0
>  %97 = shufflevector <8 x float> %96, <8 x float> undef, <8 x i32> zeroinitializer
>  %98 = getelementptr float, float* %"fs_constants[]44", i32 29
>  %99 = load float, float* %98
>  %100 = insertelement <8 x float> undef, float %99, i32 0
>  %101 = shufflevector <8 x float> %100, <8 x float> undef, <8 x i32> zeroinitializer
>  %102 = fmul <8 x float> %vY, %97
>  %103 = fadd <8 x float> %102, %101
>  store <8 x float> %103, <8 x float>* %temp47
>  %104 = fcmp oge <8 x float> %vFrontFace.splat, zeroinitializer
>  %105 = sext <8 x i1> %104 to <8 x i32>
>  %106 = bitcast <8 x i32> %105 to <8 x float>
>  store <8 x float> %106, <8 x float>* %temp50
>  store <8 x float> <float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000>, <8 x float>* %temp51
>  store <8 x float> <float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000>, <8 x float>* %temp52
>  store <8 x float> <float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000>, <8 x float>* %temp53
>  store <8 x float> %60, <8 x float>* %temp54
>  store <8 x float> %74, <8 x float>* %temp55
>  store <8 x float> %92, <8 x float>* %temp56
>  %107 = load <8 x float>, <8 x float>* %temp
>  %108 = fptosi <8 x float> %107 to <8 x i32>
>  %109 = load <8 x float>, <8 x float>* %temp47
>  %110 = fptosi <8 x float> %109 to <8 x i32>
>  %111 = bitcast <8 x i32> %108 to <8 x float>
>  store <8 x float> %111, <8 x float>* %temp58
>  %112 = bitcast <8 x i32> %110 to <8 x float>
>  store <8 x float> %112, <8 x float>* %temp59
>  %113 = load <8 x float>, <8 x float>* %temp58
>  %114 = load <8 x float>, <8 x float>* %temp59
>  store <8 x float> %113, <8 x float>* %temp58
>  store <8 x float> %114, <8 x float>* %temp59
>  store <8 x float> zeroinitializer, <8 x float>* %temp61
>  %115 = load <8 x float>, <8 x float>* %temp61
>  %116 = bitcast <8 x float> %115 to <8 x i32>
>  %117 = load <8 x float>, <8 x float>* %temp58
>  %118 = bitcast <8 x float> %117 to <8 x i32>
>  %119 = load <8 x float>, <8 x float>* %temp59
>  %120 = bitcast <8 x float> %119 to <8 x i32>
>  %121 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 0
>  %context.texture1.width = load i32, i32* %121
>  %context.texture1.row_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 6
>  %context.texture1.img_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 7
>  %122 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 5
>  %context.texture1.base_ptr = load i8*, i8** %122
>  %context.texture1.mip_offsets = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 8
>  %123 = insertelement <4 x i32> undef, i32 %context.texture1.width, i32 0
>  %124 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 1
>  %context.texture1.height = load i32, i32* %124
>  %125 = insertelement <4 x i32> %123, i32 %context.texture1.height, i32 1
>  %126 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 3
>  %context.texture1.first_level = load i32, i32* %126
>  %127 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 1, i32 4
>  %context.texture1.last_level = load i32, i32* %127
>  %128 = insertelement <8 x i32> undef, i32 %context.texture1.first_level, i32 0
>  %129 = shufflevector <8 x i32> %128, <8 x i32> undef, <8 x i32> zeroinitializer
>  %130 = insertelement <8 x i32> undef, i32 %context.texture1.last_level, i32 0
>  %131 = shufflevector <8 x i32> %130, <8 x i32> undef, <8 x i32> zeroinitializer
>  %132 = add <8 x i32> %116, %129
>  %133 = icmp slt <8 x i32> %132, %129
>  %134 = sext <8 x i1> %133 to <8 x i32>
>  %135 = icmp sgt <8 x i32> %132, %131
>  %136 = sext <8 x i1> %135 to <8 x i32>
>  %137 = or <8 x i32> %134, %136
>  %138 = xor <8 x i32> %137, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %139 = and <8 x i32> %132, %138
>  %140 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> zeroinitializer
>  %minify = lshr <4 x i32> %125, %140
>  %141 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %142 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
>  %minify74 = lshr <4 x i32> %125, %142
>  %143 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify74, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %144 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 2, i32 2, i32 2, i32 2>
>  %minify75 = lshr <4 x i32> %125, %144
>  %145 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify75, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %146 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
>  %minify76 = lshr <4 x i32> %125, %146
>  %147 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify76, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %148 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 4, i32 4, i32 4, i32 4>
>  %minify77 = lshr <4 x i32> %125, %148
>  %149 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify77, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %150 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 5, i32 5, i32 5, i32 5>
>  %minify78 = lshr <4 x i32> %125, %150
>  %151 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify78, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %152 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 6, i32 6, i32 6, i32 6>
>  %minify79 = lshr <4 x i32> %125, %152
>  %153 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify79, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %154 = shufflevector <8 x i32> %139, <8 x i32> undef, <4 x i32> <i32 7, i32 7, i32 7, i32 7>
>  %minify80 = lshr <4 x i32> %125, %154
>  %155 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify80, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %156 = shufflevector <4 x i32> %141, <4 x i32> %143, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %157 = shufflevector <4 x i32> %145, <4 x i32> %147, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %158 = shufflevector <4 x i32> %149, <4 x i32> %151, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %159 = shufflevector <4 x i32> %153, <4 x i32> %155, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %160 = shufflevector <8 x i32> %156, <8 x i32> %157, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
>  %161 = shufflevector <8 x i32> %158, <8 x i32> %159, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
>  %162 = shufflevector <16 x i32> %160, <16 x i32> %161, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
>  %163 = extractelement <8 x i32> %139, i32 0
>  %164 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %163
>  %165 = load i32, i32* %164
>  %166 = insertelement <8 x i32> undef, i32 %165, i32 0
>  %167 = extractelement <8 x i32> %139, i32 1
>  %168 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %167
>  %169 = load i32, i32* %168
>  %170 = insertelement <8 x i32> %166, i32 %169, i32 1
>  %171 = extractelement <8 x i32> %139, i32 2
>  %172 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %171
>  %173 = load i32, i32* %172
>  %174 = insertelement <8 x i32> %170, i32 %173, i32 2
>  %175 = extractelement <8 x i32> %139, i32 3
>  %176 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %175
>  %177 = load i32, i32* %176
>  %178 = insertelement <8 x i32> %174, i32 %177, i32 3
>  %179 = extractelement <8 x i32> %139, i32 4
>  %180 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %179
>  %181 = load i32, i32* %180
>  %182 = insertelement <8 x i32> %178, i32 %181, i32 4
>  %183 = extractelement <8 x i32> %139, i32 5
>  %184 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %183
>  %185 = load i32, i32* %184
>  %186 = insertelement <8 x i32> %182, i32 %185, i32 5
>  %187 = extractelement <8 x i32> %139, i32 6
>  %188 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %187
>  %189 = load i32, i32* %188
>  %190 = insertelement <8 x i32> %186, i32 %189, i32 6
>  %191 = extractelement <8 x i32> %139, i32 7
>  %192 = getelementptr [16 x i32], [16 x i32]* %context.texture1.row_stride, i32 0, i32 %191
>  %193 = load i32, i32* %192
>  %194 = insertelement <8 x i32> %190, i32 %193, i32 7
>  %195 = shufflevector <32 x i32> %162, <32 x i32> %162, <8 x i32> <i32 0, i32 4, i32 8, i32 12, i32 16, i32 20, i32 24, i32 28>
>  %196 = shufflevector <32 x i32> %162, <32 x i32> %162, <8 x i32> <i32 1, i32 5, i32 9, i32 13, i32 17, i32 21, i32 25, i32 29>
>  %197 = icmp slt <8 x i32> %118, zeroinitializer
>  %198 = sext <8 x i1> %197 to <8 x i32>
>  %199 = or <8 x i32> %137, %198
>  %200 = icmp sge <8 x i32> %118, %195
>  %201 = sext <8 x i1> %200 to <8 x i32>
>  %202 = or <8 x i32> %199, %201
>  %203 = icmp slt <8 x i32> %120, zeroinitializer
>  %204 = sext <8 x i1> %203 to <8 x i32>
>  %205 = or <8 x i32> %202, %204
>  %206 = icmp sge <8 x i32> %120, %196
>  %207 = sext <8 x i1> %206 to <8 x i32>
>  %208 = or <8 x i32> %205, %207
>  %209 = mul <8 x i32> %118, <i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8>
>  %210 = mul <8 x i32> %120, %194
>  %211 = add <8 x i32> %209, %210
>  %212 = extractelement <8 x i32> %139, i32 0
>  %213 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %212
>  %214 = load i32, i32* %213
>  %215 = insertelement <8 x i32> undef, i32 %214, i32 0
>  %216 = extractelement <8 x i32> %139, i32 1
>  %217 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %216
>  %218 = load i32, i32* %217
>  %219 = insertelement <8 x i32> %215, i32 %218, i32 1
>  %220 = extractelement <8 x i32> %139, i32 2
>  %221 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %220
>  %222 = load i32, i32* %221
>  %223 = insertelement <8 x i32> %219, i32 %222, i32 2
>  %224 = extractelement <8 x i32> %139, i32 3
>  %225 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %224
>  %226 = load i32, i32* %225
>  %227 = insertelement <8 x i32> %223, i32 %226, i32 3
>  %228 = extractelement <8 x i32> %139, i32 4
>  %229 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %228
>  %230 = load i32, i32* %229
>  %231 = insertelement <8 x i32> %227, i32 %230, i32 4
>  %232 = extractelement <8 x i32> %139, i32 5
>  %233 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %232
>  %234 = load i32, i32* %233
>  %235 = insertelement <8 x i32> %231, i32 %234, i32 5
>  %236 = extractelement <8 x i32> %139, i32 6
>  %237 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %236
>  %238 = load i32, i32* %237
>  %239 = insertelement <8 x i32> %235, i32 %238, i32 6
>  %240 = extractelement <8 x i32> %139, i32 7
>  %241 = getelementptr [16 x i32], [16 x i32]* %context.texture1.mip_offsets, i32 0, i32 %240
>  %242 = load i32, i32* %241
>  %243 = insertelement <8 x i32> %239, i32 %242, i32 7
>  %244 = add <8 x i32> %211, %243
>  %245 = xor <8 x i32> %208, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %246 = and <8 x i32> %244, %245
>  %247 = extractelement <8 x i32> %246, i32 0
>  %248 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %247
>  %249 = bitcast i8* %248 to i64*
>  %250 = load i64, i64* %249
>  %251 = zext i64 %250 to i128
>  %252 = bitcast i128 %251 to <4 x float>
>  %253 = shufflevector <4 x float> %252, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %254 = extractelement <4 x float> %253, i32 0
>  %255 = insertelement <8 x float> undef, float %254, i32 0
>  %256 = extractelement <4 x float> %253, i32 1
>  %257 = insertelement <8 x float> undef, float %256, i32 0
>  %258 = extractelement <4 x float> %253, i32 2
>  %259 = insertelement <8 x float> undef, float %258, i32 0
>  %260 = extractelement <4 x float> %253, i32 3
>  %261 = insertelement <8 x float> undef, float %260, i32 0
>  %262 = extractelement <8 x i32> %246, i32 1
>  %263 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %262
>  %264 = bitcast i8* %263 to i64*
>  %265 = load i64, i64* %264
>  %266 = zext i64 %265 to i128
>  %267 = bitcast i128 %266 to <4 x float>
>  %268 = shufflevector <4 x float> %267, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %269 = extractelement <4 x float> %268, i32 0
>  %270 = insertelement <8 x float> %255, float %269, i32 1
>  %271 = extractelement <4 x float> %268, i32 1
>  %272 = insertelement <8 x float> %257, float %271, i32 1
>  %273 = extractelement <4 x float> %268, i32 2
>  %274 = insertelement <8 x float> %259, float %273, i32 1
>  %275 = extractelement <4 x float> %268, i32 3
>  %276 = insertelement <8 x float> %261, float %275, i32 1
>  %277 = extractelement <8 x i32> %246, i32 2
>  %278 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %277
>  %279 = bitcast i8* %278 to i64*
>  %280 = load i64, i64* %279
>  %281 = zext i64 %280 to i128
>  %282 = bitcast i128 %281 to <4 x float>
>  %283 = shufflevector <4 x float> %282, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %284 = extractelement <4 x float> %283, i32 0
>  %285 = insertelement <8 x float> %270, float %284, i32 2
>  %286 = extractelement <4 x float> %283, i32 1
>  %287 = insertelement <8 x float> %272, float %286, i32 2
>  %288 = extractelement <4 x float> %283, i32 2
>  %289 = insertelement <8 x float> %274, float %288, i32 2
>  %290 = extractelement <4 x float> %283, i32 3
>  %291 = insertelement <8 x float> %276, float %290, i32 2
>  %292 = extractelement <8 x i32> %246, i32 3
>  %293 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %292
>  %294 = bitcast i8* %293 to i64*
>  %295 = load i64, i64* %294
>  %296 = zext i64 %295 to i128
>  %297 = bitcast i128 %296 to <4 x float>
>  %298 = shufflevector <4 x float> %297, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %299 = extractelement <4 x float> %298, i32 0
>  %300 = insertelement <8 x float> %285, float %299, i32 3
>  %301 = extractelement <4 x float> %298, i32 1
>  %302 = insertelement <8 x float> %287, float %301, i32 3
>  %303 = extractelement <4 x float> %298, i32 2
>  %304 = insertelement <8 x float> %289, float %303, i32 3
>  %305 = extractelement <4 x float> %298, i32 3
>  %306 = insertelement <8 x float> %291, float %305, i32 3
>  %307 = extractelement <8 x i32> %246, i32 4
>  %308 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %307
>  %309 = bitcast i8* %308 to i64*
>  %310 = load i64, i64* %309
>  %311 = zext i64 %310 to i128
>  %312 = bitcast i128 %311 to <4 x float>
>  %313 = shufflevector <4 x float> %312, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %314 = extractelement <4 x float> %313, i32 0
>  %315 = insertelement <8 x float> %300, float %314, i32 4
>  %316 = extractelement <4 x float> %313, i32 1
>  %317 = insertelement <8 x float> %302, float %316, i32 4
>  %318 = extractelement <4 x float> %313, i32 2
>  %319 = insertelement <8 x float> %304, float %318, i32 4
>  %320 = extractelement <4 x float> %313, i32 3
>  %321 = insertelement <8 x float> %306, float %320, i32 4
>  %322 = extractelement <8 x i32> %246, i32 5
>  %323 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %322
>  %324 = bitcast i8* %323 to i64*
>  %325 = load i64, i64* %324
>  %326 = zext i64 %325 to i128
>  %327 = bitcast i128 %326 to <4 x float>
>  %328 = shufflevector <4 x float> %327, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %329 = extractelement <4 x float> %328, i32 0
>  %330 = insertelement <8 x float> %315, float %329, i32 5
>  %331 = extractelement <4 x float> %328, i32 1
>  %332 = insertelement <8 x float> %317, float %331, i32 5
>  %333 = extractelement <4 x float> %328, i32 2
>  %334 = insertelement <8 x float> %319, float %333, i32 5
>  %335 = extractelement <4 x float> %328, i32 3
>  %336 = insertelement <8 x float> %321, float %335, i32 5
>  %337 = extractelement <8 x i32> %246, i32 6
>  %338 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %337
>  %339 = bitcast i8* %338 to i64*
>  %340 = load i64, i64* %339
>  %341 = zext i64 %340 to i128
>  %342 = bitcast i128 %341 to <4 x float>
>  %343 = shufflevector <4 x float> %342, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %344 = extractelement <4 x float> %343, i32 0
>  %345 = insertelement <8 x float> %330, float %344, i32 6
>  %346 = extractelement <4 x float> %343, i32 1
>  %347 = insertelement <8 x float> %332, float %346, i32 6
>  %348 = extractelement <4 x float> %343, i32 2
>  %349 = insertelement <8 x float> %334, float %348, i32 6
>  %350 = extractelement <4 x float> %343, i32 3
>  %351 = insertelement <8 x float> %336, float %350, i32 6
>  %352 = extractelement <8 x i32> %246, i32 7
>  %353 = getelementptr i8, i8* %context.texture1.base_ptr, i32 %352
>  %354 = bitcast i8* %353 to i64*
>  %355 = load i64, i64* %354
>  %356 = zext i64 %355 to i128
>  %357 = bitcast i128 %356 to <4 x float>
>  %358 = shufflevector <4 x float> %357, <4 x float> <float 0.000000e+00, float 1.000000e+00, float undef, float undef>, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
>  %359 = extractelement <4 x float> %358, i32 0
>  %360 = insertelement <8 x float> %345, float %359, i32 7
>  %361 = extractelement <4 x float> %358, i32 1
>  %362 = insertelement <8 x float> %347, float %361, i32 7
>  %363 = extractelement <4 x float> %358, i32 2
>  %364 = insertelement <8 x float> %349, float %363, i32 7
>  %365 = extractelement <4 x float> %358, i32 3
>  %366 = insertelement <8 x float> %351, float %365, i32 7
>  %367 = bitcast <8 x float> %360 to <8 x i32>
>  %368 = and <8 x i32> zeroinitializer, %208
>  %369 = xor <8 x i32> %208, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %370 = and <8 x i32> %367, %369
>  %371 = or <8 x i32> %368, %370
>  %372 = bitcast <8 x i32> %371 to <8 x float>
>  %373 = bitcast <8 x float> %362 to <8 x i32>
>  %374 = and <8 x i32> zeroinitializer, %208
>  %375 = xor <8 x i32> %208, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %376 = and <8 x i32> %373, %375
>  %377 = or <8 x i32> %374, %376
>  %378 = bitcast <8 x i32> %377 to <8 x float>
>  %379 = bitcast <8 x float> %364 to <8 x i32>
>  %380 = and <8 x i32> zeroinitializer, %208
>  %381 = xor <8 x i32> %208, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %382 = and <8 x i32> %379, %381
>  %383 = or <8 x i32> %380, %382
>  %384 = bitcast <8 x i32> %383 to <8 x float>
>  %385 = bitcast <8 x float> %366 to <8 x i32>
>  %386 = and <8 x i32> zeroinitializer, %208
>  %387 = xor <8 x i32> %208, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %388 = and <8 x i32> %385, %387
>  %389 = or <8 x i32> %386, %388
>  %390 = bitcast <8 x i32> %389 to <8 x float>
>  store <8 x float> %378, <8 x float>* %temp59
>  %391 = load <8 x float>, <8 x float>* %temp59
>  %392 = fcmp une <8 x float> %391, <float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00, float -1.000000e+00>
>  %393 = sext <8 x i1> %392 to <8 x i32>
>  %394 = bitcast <8 x i32> %393 to <8 x float>
>  store <8 x float> %394, <8 x float>* %temp62
>  %395 = load <8 x float>, <8 x float>* %temp59
>  %396 = load <8 x float>, <8 x float>* %temp48
>  %397 = fcmp olt <8 x float> %395, %396
>  %398 = sext <8 x i1> %397 to <8 x i32>
>  %399 = bitcast <8 x i32> %398 to <8 x float>
>  store <8 x float> %399, <8 x float>* %temp66
>  %400 = load <8 x float>, <8 x float>* %temp62
>  %401 = bitcast <8 x float> %400 to <8 x i32>
>  %402 = load <8 x float>, <8 x float>* %temp66
>  %403 = bitcast <8 x float> %402 to <8 x i32>
>  %404 = and <8 x i32> %401, %403
>  %405 = bitcast <8 x i32> %404 to <8 x float>
>  store <8 x float> %405, <8 x float>* %temp62
>  %406 = load <8 x float>, <8 x float>* %temp62
>  %407 = bitcast <8 x float> %406 to <8 x i32>
>  %408 = icmp ne <8 x i32> %407, zeroinitializer
>  %409 = sext <8 x i1> %408 to <8 x i32>
>  %410 = and <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, %409
>  %411 = load <8 x float>, <8 x float>* %temp62
>  %412 = bitcast <8 x float> %411 to <8 x i32>
>  %413 = and <8 x i32> <i32 -1082130432, i32 -1082130432, i32 -1082130432, i32 -1082130432, i32 -1082130432, i32 -1082130432, i32 -1082130432, i32 -1082130432>, %410
>  %414 = xor <8 x i32> %410, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %415 = and <8 x i32> %412, %414
>  %416 = or <8 x i32> %413, %415
>  %417 = bitcast <8 x i32> %416 to <8 x float>
>  store <8 x float> %417, <8 x float>* %temp62
>  %418 = load <8 x float>, <8 x float>* %temp59
>  %419 = load <8 x float>, <8 x float>* %temp63
>  %420 = bitcast <8 x i32> %410 to <8 x float>
>  %421 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %419, <8 x float> %418, <8 x float> %420)
>  store <8 x float> %421, <8 x float>* %temp63
>  %422 = load <8 x float>, <8 x float>* %temp62
>  %423 = load <8 x float>, <8 x float>* %temp63
>  %424 = load <8 x float>, <8 x float>* %temp58
>  %425 = bitcast <8 x i32> %410 to <8 x float>
>  %426 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %424, <8 x float> %422, <8 x float> %425)
>  store <8 x float> %426, <8 x float>* %temp58
>  %427 = load <8 x float>, <8 x float>* %temp59
>  %428 = bitcast <8 x i32> %410 to <8 x float>
>  %429 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %427, <8 x float> %423, <8 x float> %428)
>  store <8 x float> %429, <8 x float>* %temp59
>  %430 = load <8 x float>, <8 x float>* %temp58
>  %431 = load <8 x float>, <8 x float>* %temp59
>  %432 = load <8 x float>, <8 x float>* %temp60
>  %433 = load <8 x float>, <8 x float>* %temp61
>  %434 = load <8 x float>, <8 x float>* %output39
>  %435 = bitcast <8 x i32> %410 to <8 x float>
>  %436 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %434, <8 x float> %430, <8 x float> %435)
>  store <8 x float> %436, <8 x float>* %output39
>  %437 = load <8 x float>, <8 x float>* %output40
>  %438 = bitcast <8 x i32> %410 to <8 x float>
>  %439 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %437, <8 x float> %431, <8 x float> %438)
>  store <8 x float> %439, <8 x float>* %output40
>  %440 = load <8 x float>, <8 x float>* %output41
>  %441 = bitcast <8 x i32> %410 to <8 x float>
>  %442 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %440, <8 x float> %432, <8 x float> %441)
>  store <8 x float> %442, <8 x float>* %output41
>  %443 = load <8 x float>, <8 x float>* %output42
>  %444 = bitcast <8 x i32> %410 to <8 x float>
>  %445 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %443, <8 x float> %433, <8 x float> %444)
>  store <8 x float> %445, <8 x float>* %output42
>  %ret = xor <8 x i32> %410, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %ret_full = and <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, %ret
>  %callmask = and <8 x i32> %410, %ret_full
>  %446 = xor <8 x i32> %410, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %447 = and <8 x i32> %446, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %callmask81 = and <8 x i32> %447, %ret_full
>  %448 = load <8 x float>, <8 x float>* %temp48
>  %449 = fcmp olt <8 x float> %448, <float 1.000000e+01, float 1.000000e+01, float 1.000000e+01, float 1.000000e+01, float 1.000000e+01, float 1.000000e+01, float 1.000000e+01, float 1.000000e+01>
>  %450 = sext <8 x i1> %449 to <8 x i32>
>  %451 = bitcast <8 x i32> %450 to <8 x float>
>  %452 = load <8 x float>, <8 x float>* %temp62
>  %453 = bitcast <8 x i32> %callmask81 to <8 x float>
>  %454 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %452, <8 x float> %451, <8 x float> %453)
>  store <8 x float> %454, <8 x float>* %temp62
>  %455 = load <8 x float>, <8 x float>* %temp62
>  %456 = bitcast <8 x float> %455 to <8 x i32>
>  %457 = icmp ne <8 x i32> %456, zeroinitializer
>  %458 = sext <8 x i1> %457 to <8 x i32>
>  %459 = and <8 x i32> %447, %458
>  %callmask82 = and <8 x i32> %459, %ret_full
>  %460 = load <8 x float>, <8 x float>* %temp48
>  %461 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %460
>  %462 = load <8 x float>, <8 x float>* %temp62
>  %463 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %464 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %462, <8 x float> %461, <8 x float> %463)
>  store <8 x float> %464, <8 x float>* %temp62
>  %465 = load <8 x float>, <8 x float>* %temp48
>  %466 = load <8 x float>, <8 x float>* %temp63
>  %467 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %468 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %466, <8 x float> %465, <8 x float> %467)
>  store <8 x float> %468, <8 x float>* %temp63
>  %469 = load <8 x float>, <8 x float>* %temp62
>  %470 = load <8 x float>, <8 x float>* %temp63
>  %471 = load <8 x float>, <8 x float>* %temp58
>  %472 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %473 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %471, <8 x float> %469, <8 x float> %472)
>  store <8 x float> %473, <8 x float>* %temp58
>  %474 = load <8 x float>, <8 x float>* %temp59
>  %475 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %476 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %474, <8 x float> %470, <8 x float> %475)
>  store <8 x float> %476, <8 x float>* %temp59
>  %477 = load <8 x float>, <8 x float>* %temp58
>  %478 = load <8 x float>, <8 x float>* %temp59
>  %479 = load <8 x float>, <8 x float>* %temp60
>  %480 = load <8 x float>, <8 x float>* %temp61
>  %481 = load <8 x float>, <8 x float>* %output39
>  %482 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %483 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %481, <8 x float> %477, <8 x float> %482)
>  store <8 x float> %483, <8 x float>* %output39
>  %484 = load <8 x float>, <8 x float>* %output40
>  %485 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %486 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %484, <8 x float> %478, <8 x float> %485)
>  store <8 x float> %486, <8 x float>* %output40
>  %487 = load <8 x float>, <8 x float>* %output41
>  %488 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %489 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %487, <8 x float> %479, <8 x float> %488)
>  store <8 x float> %489, <8 x float>* %output41
>  %490 = load <8 x float>, <8 x float>* %output42
>  %491 = bitcast <8 x i32> %callmask82 to <8 x float>
>  %492 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %490, <8 x float> %480, <8 x float> %491)
>  store <8 x float> %492, <8 x float>* %output42
>  %ret83 = xor <8 x i32> %callmask82, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %ret_full84 = and <8 x i32> %ret_full, %ret83
>  %callmask85 = and <8 x i32> %459, %ret_full84
>  %callmask86 = and <8 x i32> %447, %ret_full84
>  %callmask87 = and <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, %ret_full84
>  %493 = load <8 x float>, <8 x float>* %temp54
>  %494 = load <8 x float>, <8 x float>* %temp55
>  %495 = load <8 x float>, <8 x float>* %temp56
>  %496 = load <8 x float>, <8 x float>* %temp54
>  %497 = load <8 x float>, <8 x float>* %temp55
>  %498 = load <8 x float>, <8 x float>* %temp56
>  %499 = fmul <8 x float> %493, %496
>  %500 = fmul <8 x float> %494, %497
>  %501 = fadd <8 x float> %500, %499
>  %502 = fmul <8 x float> %495, %498
>  %503 = fadd <8 x float> %501, %502
>  %504 = load <8 x float>, <8 x float>* %temp62
>  %505 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %506 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %504, <8 x float> %503, <8 x float> %505)
>  store <8 x float> %506, <8 x float>* %temp62
>  %507 = load <8 x float>, <8 x float>* %temp62
>  %508 = call <8 x float> @llvm.sqrt.v8f32(<8 x float> %507)
>  %509 = fdiv <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %508
>  %510 = load <8 x float>, <8 x float>* %temp62
>  %511 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %512 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %510, <8 x float> %509, <8 x float> %511)
>  store <8 x float> %512, <8 x float>* %temp62
>  %513 = load <8 x float>, <8 x float>* %temp54
>  %514 = load <8 x float>, <8 x float>* %temp62
>  %515 = fmul <8 x float> %513, %514
>  %516 = load <8 x float>, <8 x float>* %temp55
>  %517 = load <8 x float>, <8 x float>* %temp62
>  %518 = fmul <8 x float> %516, %517
>  %519 = load <8 x float>, <8 x float>* %temp56
>  %520 = load <8 x float>, <8 x float>* %temp62
>  %521 = fmul <8 x float> %519, %520
>  %522 = load <8 x float>, <8 x float>* %temp62
>  %523 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %524 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %522, <8 x float> %515, <8 x float> %523)
>  store <8 x float> %524, <8 x float>* %temp62
>  %525 = load <8 x float>, <8 x float>* %temp63
>  %526 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %527 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %525, <8 x float> %518, <8 x float> %526)
>  store <8 x float> %527, <8 x float>* %temp63
>  %528 = load <8 x float>, <8 x float>* %temp64
>  %529 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %530 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %528, <8 x float> %521, <8 x float> %529)
>  store <8 x float> %530, <8 x float>* %temp64
>  %531 = load <8 x float>, <8 x float>* %temp64
>  %532 = load <8 x float>, <8 x float>* %temp68
>  %533 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %534 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %532, <8 x float> %531, <8 x float> %533)
>  store <8 x float> %534, <8 x float>* %temp68
>  %535 = load <8 x float>, <8 x float>* %temp50
>  %536 = bitcast <8 x float> %535 to <8 x i32>
>  %537 = icmp eq <8 x i32> %536, zeroinitializer
>  %538 = sext <8 x i1> %537 to <8 x i32>
>  %539 = bitcast <8 x i32> %538 to <8 x float>
>  %540 = load <8 x float>, <8 x float>* %temp70
>  %541 = bitcast <8 x i32> %callmask87 to <8 x float>
>  %542 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %540, <8 x float> %539, <8 x float> %541)
>  store <8 x float> %542, <8 x float>* %temp70
>  %543 = load <8 x float>, <8 x float>* %temp70
>  %544 = bitcast <8 x float> %543 to <8 x i32>
>  %545 = icmp ne <8 x i32> %544, zeroinitializer
>  %546 = sext <8 x i1> %545 to <8 x i32>
>  %547 = and <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, %546
>  %callmask88 = and <8 x i32> %547, %ret_full84
>  %548 = load <8 x float>, <8 x float>* %temp64
>  %549 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %548
>  %550 = load <8 x float>, <8 x float>* %temp68
>  %551 = bitcast <8 x i32> %callmask88 to <8 x float>
>  %552 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %550, <8 x float> %549, <8 x float> %551)
>  store <8 x float> %552, <8 x float>* %temp68
>  %callmask89 = and <8 x i32> <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, %ret_full84
>  %553 = load <8 x float>, <8 x float>* %temp68
>  %554 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> zeroinitializer, <8 x float> %553)
>  %isnotnan = fcmp oeq <8 x float> %553, %553
>  %555 = xor <8 x i1> %isnotnan, <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>
>  %isnan = sext <8 x i1> %555 to <8 x i32>
>  %556 = bitcast <8 x float> %554 to <8 x i32>
>  %557 = and <8 x i32> zeroinitializer, %isnan
>  %558 = xor <8 x i32> %isnan, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %559 = and <8 x i32> %556, %558
>  %560 = or <8 x i32> %557, %559
>  %561 = bitcast <8 x i32> %560 to <8 x float>
>  %562 = load <8 x float>, <8 x float>* %temp62
>  %563 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %564 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %562, <8 x float> %561, <8 x float> %563)
>  store <8 x float> %564, <8 x float>* %temp62
>  %565 = load <8 x float>, <8 x float>* %temp62
>  %566 = getelementptr float, float* %"fs_constants[]44", i32 16
>  %567 = load float, float* %566
>  %568 = insertelement <8 x float> undef, float %567, i32 0
>  %569 = shufflevector <8 x float> %568, <8 x float> undef, <8 x i32> zeroinitializer
>  %570 = bitcast <8 x float> %565 to <8 x i32>
>  %571 = and <8 x i32> %570, <i32 2139095040, i32 2139095040, i32 2139095040, i32 2139095040, i32 2139095040, i32 2139095040, i32 2139095040, i32 2139095040>
>  %572 = lshr <8 x i32> %571, <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23>
>  %573 = sub <8 x i32> %572, <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127>
>  %574 = sitofp <8 x i32> %573 to <8 x float>
>  %575 = and <8 x i32> %570, <i32 8388607, i32 8388607, i32 8388607, i32 8388607, i32 8388607, i32 8388607, i32 8388607, i32 8388607>
>  %576 = or <8 x i32> %575, <i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216>
>  %577 = bitcast <8 x i32> %576 to <8 x float>
>  %578 = fadd <8 x float> %577, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
>  %579 = fsub <8 x float> %577, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
>  %580 = fdiv <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %578
>  %581 = fmul <8 x float> %579, %580
>  %582 = fmul <8 x float> %581, %581
>  %583 = fmul <8 x float> %582, %582
>  %584 = fmul <8 x float> %583, <float 0x3FDA07AB20000000, float 0x3FDA07AB20000000, float 0x3FDA07AB20000000, float 0x3FDA07AB20000000, float 0x3FDA07AB20000000, float 0x3FDA07AB20000000, float 0x3FDA07AB20000000, float 0x3FDA07AB20000000>
>  %585 = fadd <8 x float> <float 0x3FE27A6420000000, float 0x3FE27A6420000000, float 0x3FE27A6420000000, float 0x3FE27A6420000000, float 0x3FE27A6420000000, float 0x3FE27A6420000000, float 0x3FE27A6420000000, float 0x3FE27A6420000000>, %584
>  %586 = fmul <8 x float> %583, <float 0x3FD9D062C0000000, float 0x3FD9D062C0000000, float 0x3FD9D062C0000000, float 0x3FD9D062C0000000, float 0x3FD9D062C0000000, float 0x3FD9D062C0000000, float 0x3FD9D062C0000000, float 0x3FD9D062C0000000>
>  %587 = fadd <8 x float> <float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000, float 0x3FEEC6FF20000000>, %586
>  %588 = fmul <8 x float> %583, %585
>  %589 = fadd <8 x float> <float 0x4007154760000000, float 0x4007154760000000, float 0x4007154760000000, float 0x4007154760000000, float 0x4007154760000000, float 0x4007154760000000, float 0x4007154760000000, float 0x4007154760000000>, %588
>  %590 = fmul <8 x float> %587, %582
>  %591 = fadd <8 x float> %590, %589
>  %592 = fmul <8 x float> %581, %591
>  %593 = fadd <8 x float> %592, %574
>  %594 = fmul <8 x float> %593, %569
>  %595 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> <float 1.280000e+02, float 1.280000e+02, float 1.280000e+02, float 1.280000e+02, float 1.280000e+02, float 1.280000e+02, float 1.280000e+02, float 1.280000e+02>, <8 x float> %594)
>  %596 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> <float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000, float 0xC05FBFFFE0000000>, <8 x float> %595)
>  %597 = call <8 x float> @llvm.floor.v8f32(<8 x float> %596)
>  %fpart = fsub <8 x float> %596, %597
>  %ipart = fptosi <8 x float> %597 to <8 x i32>
>  %598 = add <8 x i32> %ipart, <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127>
>  %599 = shl <8 x i32> %598, <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23>
>  %600 = bitcast <8 x i32> %599 to <8 x float>
>  %601 = fmul <8 x float> %fpart, %fpart
>  %602 = fmul <8 x float> %601, <float 0x3F5EC320A0000000, float 0x3F5EC320A0000000, float 0x3F5EC320A0000000, float 0x3F5EC320A0000000, float 0x3F5EC320A0000000, float 0x3F5EC320A0000000, float 0x3F5EC320A0000000, float 0x3F5EC320A0000000>
>  %603 = fadd <8 x float> <float 0x3FAC954460000000, float 0x3FAC954460000000, float 0x3FAC954460000000, float 0x3FAC954460000000, float 0x3FAC954460000000, float 0x3FAC954460000000, float 0x3FAC954460000000, float 0x3FAC954460000000>, %602
>  %604 = fmul <8 x float> %601, <float 0x3F826900C0000000, float 0x3F826900C0000000, float 0x3F826900C0000000, float 0x3F826900C0000000, float 0x3F826900C0000000, float 0x3F826900C0000000, float 0x3F826900C0000000, float 0x3F826900C0000000>
>  %605 = fadd <8 x float> <float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000, float 0x3FCEBD5A80000000>, %604
>  %606 = fmul <8 x float> %601, %603
>  %607 = fadd <8 x float> <float 0x3FE62E4F60000000, float 0x3FE62E4F60000000, float 0x3FE62E4F60000000, float 0x3FE62E4F60000000, float 0x3FE62E4F60000000, float 0x3FE62E4F60000000, float 0x3FE62E4F60000000, float 0x3FE62E4F60000000>, %606
>  %608 = fmul <8 x float> %601, %605
>  %609 = fadd <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, %608
>  %610 = fmul <8 x float> %607, %fpart
>  %611 = fadd <8 x float> %610, %609
>  %612 = fmul <8 x float> %600, %611
>  %613 = load <8 x float>, <8 x float>* %temp66
>  %614 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %615 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %613, <8 x float> %612, <8 x float> %614)
>  store <8 x float> %615, <8 x float>* %temp66
>  %616 = load <8 x float>, <8 x float>* %temp62
>  %617 = getelementptr float, float* %"fs_constants[]44", i32 8
>  %618 = load float, float* %617
>  %619 = insertelement <8 x float> undef, float %618, i32 0
>  %620 = shufflevector <8 x float> %619, <8 x float> undef, <8 x i32> zeroinitializer
>  %621 = getelementptr float, float* %"fs_constants[]44", i32 4
>  %622 = load float, float* %621
>  %623 = insertelement <8 x float> undef, float %622, i32 0
>  %624 = shufflevector <8 x float> %623, <8 x float> undef, <8 x i32> zeroinitializer
>  %625 = fmul <8 x float> %616, %620
>  %626 = fadd <8 x float> %625, %624
>  %627 = load <8 x float>, <8 x float>* %temp62
>  %628 = getelementptr float, float* %"fs_constants[]44", i32 9
>  %629 = load float, float* %628
>  %630 = insertelement <8 x float> undef, float %629, i32 0
>  %631 = shufflevector <8 x float> %630, <8 x float> undef, <8 x i32> zeroinitializer
>  %632 = getelementptr float, float* %"fs_constants[]44", i32 5
>  %633 = load float, float* %632
>  %634 = insertelement <8 x float> undef, float %633, i32 0
>  %635 = shufflevector <8 x float> %634, <8 x float> undef, <8 x i32> zeroinitializer
>  %636 = fmul <8 x float> %627, %631
>  %637 = fadd <8 x float> %636, %635
>  %638 = load <8 x float>, <8 x float>* %temp62
>  %639 = getelementptr float, float* %"fs_constants[]44", i32 10
>  %640 = load float, float* %639
>  %641 = insertelement <8 x float> undef, float %640, i32 0
>  %642 = shufflevector <8 x float> %641, <8 x float> undef, <8 x i32> zeroinitializer
>  %643 = getelementptr float, float* %"fs_constants[]44", i32 6
>  %644 = load float, float* %643
>  %645 = insertelement <8 x float> undef, float %644, i32 0
>  %646 = shufflevector <8 x float> %645, <8 x float> undef, <8 x i32> zeroinitializer
>  %647 = fmul <8 x float> %638, %642
>  %648 = fadd <8 x float> %647, %646
>  %649 = load <8 x float>, <8 x float>* %temp62
>  %650 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %651 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %649, <8 x float> %626, <8 x float> %650)
>  store <8 x float> %651, <8 x float>* %temp62
>  %652 = load <8 x float>, <8 x float>* %temp63
>  %653 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %654 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %652, <8 x float> %637, <8 x float> %653)
>  store <8 x float> %654, <8 x float>* %temp63
>  %655 = load <8 x float>, <8 x float>* %temp64
>  %656 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %657 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %655, <8 x float> %648, <8 x float> %656)
>  store <8 x float> %657, <8 x float>* %temp64
>  %658 = load <8 x float>, <8 x float>* %temp66
>  %659 = getelementptr float, float* %"fs_constants[]44", i32 12
>  %660 = load float, float* %659
>  %661 = insertelement <8 x float> undef, float %660, i32 0
>  %662 = shufflevector <8 x float> %661, <8 x float> undef, <8 x i32> zeroinitializer
>  %663 = load <8 x float>, <8 x float>* %temp62
>  %664 = fmul <8 x float> %658, %662
>  %665 = fadd <8 x float> %664, %663
>  %666 = load <8 x float>, <8 x float>* %temp66
>  %667 = getelementptr float, float* %"fs_constants[]44", i32 13
>  %668 = load float, float* %667
>  %669 = insertelement <8 x float> undef, float %668, i32 0
>  %670 = shufflevector <8 x float> %669, <8 x float> undef, <8 x i32> zeroinitializer
>  %671 = load <8 x float>, <8 x float>* %temp63
>  %672 = fmul <8 x float> %666, %670
>  %673 = fadd <8 x float> %672, %671
>  %674 = load <8 x float>, <8 x float>* %temp66
>  %675 = getelementptr float, float* %"fs_constants[]44", i32 14
>  %676 = load float, float* %675
>  %677 = insertelement <8 x float> undef, float %676, i32 0
>  %678 = shufflevector <8 x float> %677, <8 x float> undef, <8 x i32> zeroinitializer
>  %679 = load <8 x float>, <8 x float>* %temp64
>  %680 = fmul <8 x float> %674, %678
>  %681 = fadd <8 x float> %680, %679
>  %682 = load <8 x float>, <8 x float>* %temp62
>  %683 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %684 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %682, <8 x float> %665, <8 x float> %683)
>  store <8 x float> %684, <8 x float>* %temp62
>  %685 = load <8 x float>, <8 x float>* %temp63
>  %686 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %687 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %685, <8 x float> %673, <8 x float> %686)
>  store <8 x float> %687, <8 x float>* %temp63
>  %688 = load <8 x float>, <8 x float>* %temp64
>  %689 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %690 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %688, <8 x float> %681, <8 x float> %689)
>  store <8 x float> %690, <8 x float>* %temp64
>  %691 = getelementptr float, float* %"fs_constants[]44", i32 0
>  %692 = load float, float* %691
>  %693 = insertelement <8 x float> undef, float %692, i32 0
>  %694 = shufflevector <8 x float> %693, <8 x float> undef, <8 x i32> zeroinitializer
>  %695 = load <8 x float>, <8 x float>* %temp65
>  %696 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %697 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %695, <8 x float> %694, <8 x float> %696)
>  store <8 x float> %697, <8 x float>* %temp65
>  %698 = load <8 x float>, <8 x float>* %temp62
>  %699 = load <8 x float>, <8 x float>* %temp63
>  %700 = load <8 x float>, <8 x float>* %temp64
>  %701 = load <8 x float>, <8 x float>* %temp65
>  %702 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %698, <8 x float> zeroinitializer)
>  %703 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %702, <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>)
>  %704 = load <8 x float>, <8 x float>* %temp62
>  %705 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %706 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %704, <8 x float> %703, <8 x float> %705)
>  store <8 x float> %706, <8 x float>* %temp62
>  %707 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %699, <8 x float> zeroinitializer)
>  %708 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %707, <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>)
>  %709 = load <8 x float>, <8 x float>* %temp63
>  %710 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %711 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %709, <8 x float> %708, <8 x float> %710)
>  store <8 x float> %711, <8 x float>* %temp63
>  %712 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %700, <8 x float> zeroinitializer)
>  %713 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %712, <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>)
>  %714 = load <8 x float>, <8 x float>* %temp64
>  %715 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %716 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %714, <8 x float> %713, <8 x float> %715)
>  store <8 x float> %716, <8 x float>* %temp64
>  %717 = call <8 x float> @llvm.x86.avx.max.ps.256(<8 x float> %701, <8 x float> zeroinitializer)
>  %718 = call <8 x float> @llvm.x86.avx.min.ps.256(<8 x float> %717, <8 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>)
>  %719 = load <8 x float>, <8 x float>* %temp65
>  %720 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %721 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %719, <8 x float> %718, <8 x float> %720)
>  store <8 x float> %721, <8 x float>* %temp65
>  %722 = load <8 x float>, <8 x float>* %temp66
>  %723 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %724 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %722, <8 x float> %32, <8 x float> %723)
>  store <8 x float> %724, <8 x float>* %temp66
>  %725 = load <8 x float>, <8 x float>* %temp67
>  %726 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %727 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %725, <8 x float> %46, <8 x float> %726)
>  store <8 x float> %727, <8 x float>* %temp67
>  %728 = load <8 x float>, <8 x float>* %temp66
>  %729 = load <8 x float>, <8 x float>* %temp67
>  %730 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 0
>  %context.texture0.width = load i32, i32* %730
>  %context.texture0.row_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 6
>  %context.texture0.img_stride = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 7
>  %731 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 5
>  %context.texture0.base_ptr = load i8*, i8** %731
>  %context.texture0.mip_offsets = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 8
>  %732 = insertelement <4 x i32> undef, i32 %context.texture0.width, i32 0
>  %733 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 1
>  %context.texture0.height = load i32, i32* %733
>  %734 = insertelement <4 x i32> %732, i32 %context.texture0.height, i32 1
>  %735 = getelementptr { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }, { [32 x float*], [32 x i32], [32 x float*], [32 x i32], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [32 x { i32, i32, i32, i32, i32, i8*, [16 x i32], [16 x i32], [16 x i32] }], [32 x { float, float, float, [4 x float] }], [10 x { i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i8, i32, i32, i32, i32, [2 x [15 x i32]], i8* }] }* %hPrivateData, i32 0, i32 6, i32 0, i32 3
>  %context.texture0.first_level = load i32, i32* %735
>  %736 = shufflevector <8 x float> %728, <8 x float> %728, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %737 = shufflevector <8 x float> %729, <8 x float> %729, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  store <16 x i8> zeroinitializer, <16 x i8>* %packed_var
>  %738 = insertelement <4 x i32> undef, i32 %context.texture0.first_level, i32 0
>  %739 = shufflevector <4 x i32> %738, <4 x i32> undef, <4 x i32> zeroinitializer
>  %minify90 = lshr <4 x i32> %734, %739
>  %740 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify90, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %741 = getelementptr [16 x i32], [16 x i32]* %context.texture0.row_stride, i32 0, i32 %context.texture0.first_level
>  %742 = load i32, i32* %741
>  %743 = insertelement <4 x i32> undef, i32 %742, i32 0
>  %744 = shufflevector <4 x i32> %743, <4 x i32> undef, <4 x i32> zeroinitializer
>  %745 = getelementptr [16 x i32], [16 x i32]* %context.texture0.mip_offsets, i32 0, i32 %context.texture0.first_level
>  %746 = load i32, i32* %745
>  %747 = getelementptr i8, i8* %context.texture0.base_ptr, i32 %746
>  %748 = shufflevector <4 x i32> %740, <4 x i32> undef, <4 x i32> zeroinitializer
>  %749 = shufflevector <4 x i32> %740, <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
>  %750 = shl <4 x i32> %740, <i32 8, i32 8, i32 8, i32 8>
>  %751 = sitofp <4 x i32> %750 to <4 x float>
>  %752 = shufflevector <4 x float> %751, <4 x float> undef, <4 x i32> zeroinitializer
>  %753 = shufflevector <4 x float> %751, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
>  %754 = fmul <4 x float> %736, %752
>  %755 = fmul <4 x float> %737, %753
>  %756 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %754)
>  %757 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %755)
>  %758 = add <4 x i32> %756, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %759 = add <4 x i32> %757, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %760 = ashr <4 x i32> %758, <i32 8, i32 8, i32 8, i32 8>
>  %761 = ashr <4 x i32> %759, <i32 8, i32 8, i32 8, i32 8>
>  %762 = and <4 x i32> %758, <i32 255, i32 255, i32 255, i32 255>
>  %763 = and <4 x i32> %759, <i32 255, i32 255, i32 255, i32 255>
>  %764 = sub <4 x i32> %748, <i32 1, i32 1, i32 1, i32 1>
>  %765 = sitofp <4 x i32> %748 to <4 x float>
>  %766 = sub <4 x i32> %748, <i32 1, i32 1, i32 1, i32 1>
>  %767 = call <4 x float> @llvm.floor.v4f32(<4 x float> %736)
>  %768 = fsub <4 x float> %736, %767
>  %769 = fmul <4 x float> %768, %765
>  %770 = fmul <4 x float> %769, <float 2.560000e+02, float 2.560000e+02, float 2.560000e+02, float 2.560000e+02>
>  %771 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %770)
>  %772 = add <4 x i32> %771, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %773 = and <4 x i32> %772, <i32 255, i32 255, i32 255, i32 255>
>  %774 = ashr <4 x i32> %772, <i32 8, i32 8, i32 8, i32 8>
>  %775 = icmp slt <4 x i32> %774, zeroinitializer
>  %776 = sext <4 x i1> %775 to <4 x i32>
>  %777 = bitcast <4 x i32> %776 to <16 x i8>
>  %778 = bitcast <4 x i32> %766 to <16 x i8>
>  %779 = bitcast <4 x i32> %774 to <16 x i8>
>  %780 = call <16 x i8> @llvm.x86.sse41.pblendvb(<16 x i8> %779, <16 x i8> %778, <16 x i8> %777)
>  %781 = bitcast <16 x i8> %780 to <4 x i32>
>  %782 = icmp ne <4 x i32> %781, %764
>  %783 = sext <4 x i1> %782 to <4 x i32>
>  %784 = mul <4 x i32> %781, <i32 4, i32 4, i32 4, i32 4>
>  %785 = add <4 x i32> %784, <i32 4, i32 4, i32 4, i32 4>
>  %786 = and <4 x i32> %785, %783
>  %787 = sub <4 x i32> %749, <i32 1, i32 1, i32 1, i32 1>
>  %788 = sitofp <4 x i32> %749 to <4 x float>
>  %789 = sub <4 x i32> %749, <i32 1, i32 1, i32 1, i32 1>
>  %790 = call <4 x float> @llvm.floor.v4f32(<4 x float> %737)
>  %791 = fsub <4 x float> %737, %790
>  %792 = fmul <4 x float> %791, %788
>  %793 = fmul <4 x float> %792, <float 2.560000e+02, float 2.560000e+02, float 2.560000e+02, float 2.560000e+02>
>  %794 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %793)
>  %795 = add <4 x i32> %794, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %796 = and <4 x i32> %795, <i32 255, i32 255, i32 255, i32 255>
>  %797 = ashr <4 x i32> %795, <i32 8, i32 8, i32 8, i32 8>
>  %798 = icmp slt <4 x i32> %797, zeroinitializer
>  %799 = sext <4 x i1> %798 to <4 x i32>
>  %800 = bitcast <4 x i32> %799 to <16 x i8>
>  %801 = bitcast <4 x i32> %789 to <16 x i8>
>  %802 = bitcast <4 x i32> %797 to <16 x i8>
>  %803 = call <16 x i8> @llvm.x86.sse41.pblendvb(<16 x i8> %802, <16 x i8> %801, <16 x i8> %800)
>  %804 = bitcast <16 x i8> %803 to <4 x i32>
>  %805 = icmp ne <4 x i32> %804, %787
>  %806 = sext <4 x i1> %805 to <4 x i32>
>  %807 = mul <4 x i32> %804, %744
>  %808 = add <4 x i32> %807, %744
>  %809 = and <4 x i32> %808, %806
>  %810 = add <4 x i32> %784, %807
>  %811 = add <4 x i32> %784, %809
>  %812 = add <4 x i32> %786, %807
>  %813 = add <4 x i32> %786, %809
>  %814 = add <4 x i32> %784, %807
>  %815 = add <4 x i32> %784, %809
>  %816 = add <4 x i32> %786, %807
>  %817 = add <4 x i32> %786, %809
>  %818 = bitcast <4 x i32> %773 to <16 x i8>
>  %819 = bitcast <4 x i32> %796 to <16 x i8>
>  %820 = shufflevector <16 x i8> %818, <16 x i8> undef, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12>
>  %821 = shufflevector <16 x i8> %819, <16 x i8> undef, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12>
>  %822 = extractelement <4 x i32> %810, i32 0
>  %823 = getelementptr i8, i8* %747, i32 %822
>  %824 = bitcast i8* %823 to i32*
>  %825 = load i32, i32* %824
>  %826 = insertelement <4 x i32> undef, i32 %825, i32 0
>  %827 = extractelement <4 x i32> %810, i32 1
>  %828 = getelementptr i8, i8* %747, i32 %827
>  %829 = bitcast i8* %828 to i32*
>  %830 = load i32, i32* %829
>  %831 = insertelement <4 x i32> %826, i32 %830, i32 1
>  %832 = extractelement <4 x i32> %810, i32 2
>  %833 = getelementptr i8, i8* %747, i32 %832
>  %834 = bitcast i8* %833 to i32*
>  %835 = load i32, i32* %834
>  %836 = insertelement <4 x i32> %831, i32 %835, i32 2
>  %837 = extractelement <4 x i32> %810, i32 3
>  %838 = getelementptr i8, i8* %747, i32 %837
>  %839 = bitcast i8* %838 to i32*
>  %840 = load i32, i32* %839
>  %841 = insertelement <4 x i32> %836, i32 %840, i32 3
>  %842 = bitcast <4 x i32> %841 to <16 x i8>
>  %843 = extractelement <4 x i32> %812, i32 0
>  %844 = getelementptr i8, i8* %747, i32 %843
>  %845 = bitcast i8* %844 to i32*
>  %846 = load i32, i32* %845
>  %847 = insertelement <4 x i32> undef, i32 %846, i32 0
>  %848 = extractelement <4 x i32> %812, i32 1
>  %849 = getelementptr i8, i8* %747, i32 %848
>  %850 = bitcast i8* %849 to i32*
>  %851 = load i32, i32* %850
>  %852 = insertelement <4 x i32> %847, i32 %851, i32 1
>  %853 = extractelement <4 x i32> %812, i32 2
>  %854 = getelementptr i8, i8* %747, i32 %853
>  %855 = bitcast i8* %854 to i32*
>  %856 = load i32, i32* %855
>  %857 = insertelement <4 x i32> %852, i32 %856, i32 2
>  %858 = extractelement <4 x i32> %812, i32 3
>  %859 = getelementptr i8, i8* %747, i32 %858
>  %860 = bitcast i8* %859 to i32*
>  %861 = load i32, i32* %860
>  %862 = insertelement <4 x i32> %857, i32 %861, i32 3
>  %863 = bitcast <4 x i32> %862 to <16 x i8>
>  %864 = extractelement <4 x i32> %811, i32 0
>  %865 = getelementptr i8, i8* %747, i32 %864
>  %866 = bitcast i8* %865 to i32*
>  %867 = load i32, i32* %866
>  %868 = insertelement <4 x i32> undef, i32 %867, i32 0
>  %869 = extractelement <4 x i32> %811, i32 1
>  %870 = getelementptr i8, i8* %747, i32 %869
>  %871 = bitcast i8* %870 to i32*
>  %872 = load i32, i32* %871
>  %873 = insertelement <4 x i32> %868, i32 %872, i32 1
>  %874 = extractelement <4 x i32> %811, i32 2
>  %875 = getelementptr i8, i8* %747, i32 %874
>  %876 = bitcast i8* %875 to i32*
>  %877 = load i32, i32* %876
>  %878 = insertelement <4 x i32> %873, i32 %877, i32 2
>  %879 = extractelement <4 x i32> %811, i32 3
>  %880 = getelementptr i8, i8* %747, i32 %879
>  %881 = bitcast i8* %880 to i32*
>  %882 = load i32, i32* %881
>  %883 = insertelement <4 x i32> %878, i32 %882, i32 3
>  %884 = bitcast <4 x i32> %883 to <16 x i8>
>  %885 = extractelement <4 x i32> %813, i32 0
>  %886 = getelementptr i8, i8* %747, i32 %885
>  %887 = bitcast i8* %886 to i32*
>  %888 = load i32, i32* %887
>  %889 = insertelement <4 x i32> undef, i32 %888, i32 0
>  %890 = extractelement <4 x i32> %813, i32 1
>  %891 = getelementptr i8, i8* %747, i32 %890
>  %892 = bitcast i8* %891 to i32*
>  %893 = load i32, i32* %892
>  %894 = insertelement <4 x i32> %889, i32 %893, i32 1
>  %895 = extractelement <4 x i32> %813, i32 2
>  %896 = getelementptr i8, i8* %747, i32 %895
>  %897 = bitcast i8* %896 to i32*
>  %898 = load i32, i32* %897
>  %899 = insertelement <4 x i32> %894, i32 %898, i32 2
>  %900 = extractelement <4 x i32> %813, i32 3
>  %901 = getelementptr i8, i8* %747, i32 %900
>  %902 = bitcast i8* %901 to i32*
>  %903 = load i32, i32* %902
>  %904 = insertelement <4 x i32> %899, i32 %903, i32 3
>  %905 = bitcast <4 x i32> %904 to <16 x i8>
>  %906 = shufflevector <16 x i8> %820, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %907 = shufflevector <16 x i8> %820, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %908 = bitcast <16 x i8> %906 to <8 x i16>
>  %909 = bitcast <16 x i8> %907 to <8 x i16>
>  %910 = shufflevector <16 x i8> %842, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %911 = shufflevector <16 x i8> %842, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %912 = bitcast <16 x i8> %910 to <8 x i16>
>  %913 = bitcast <16 x i8> %911 to <8 x i16>
>  %914 = shufflevector <16 x i8> %863, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %915 = shufflevector <16 x i8> %863, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %916 = bitcast <16 x i8> %914 to <8 x i16>
>  %917 = bitcast <16 x i8> %915 to <8 x i16>
>  %918 = sub <8 x i16> %916, %912
>  %919 = mul <8 x i16> %908, %918
>  %920 = lshr <8 x i16> %919, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %921 = add <8 x i16> %912, %920
>  %922 = and <8 x i16> %921, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %923 = sub <8 x i16> %917, %913
>  %924 = mul <8 x i16> %909, %923
>  %925 = lshr <8 x i16> %924, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %926 = add <8 x i16> %913, %925
>  %927 = and <8 x i16> %926, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %928 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %922, <8 x i16> %927)
>  %929 = shufflevector <16 x i8> %820, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %930 = shufflevector <16 x i8> %820, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %931 = bitcast <16 x i8> %929 to <8 x i16>
>  %932 = bitcast <16 x i8> %930 to <8 x i16>
>  %933 = shufflevector <16 x i8> %884, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %934 = shufflevector <16 x i8> %884, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %935 = bitcast <16 x i8> %933 to <8 x i16>
>  %936 = bitcast <16 x i8> %934 to <8 x i16>
>  %937 = shufflevector <16 x i8> %905, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %938 = shufflevector <16 x i8> %905, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %939 = bitcast <16 x i8> %937 to <8 x i16>
>  %940 = bitcast <16 x i8> %938 to <8 x i16>
>  %941 = sub <8 x i16> %939, %935
>  %942 = mul <8 x i16> %931, %941
>  %943 = lshr <8 x i16> %942, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %944 = add <8 x i16> %935, %943
>  %945 = and <8 x i16> %944, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %946 = sub <8 x i16> %940, %936
>  %947 = mul <8 x i16> %932, %946
>  %948 = lshr <8 x i16> %947, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %949 = add <8 x i16> %936, %948
>  %950 = and <8 x i16> %949, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %951 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %945, <8 x i16> %950)
>  %952 = shufflevector <16 x i8> %821, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %953 = shufflevector <16 x i8> %821, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %954 = bitcast <16 x i8> %952 to <8 x i16>
>  %955 = bitcast <16 x i8> %953 to <8 x i16>
>  %956 = shufflevector <16 x i8> %928, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %957 = shufflevector <16 x i8> %928, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %958 = bitcast <16 x i8> %956 to <8 x i16>
>  %959 = bitcast <16 x i8> %957 to <8 x i16>
>  %960 = shufflevector <16 x i8> %951, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %961 = shufflevector <16 x i8> %951, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %962 = bitcast <16 x i8> %960 to <8 x i16>
>  %963 = bitcast <16 x i8> %961 to <8 x i16>
>  %964 = sub <8 x i16> %962, %958
>  %965 = mul <8 x i16> %954, %964
>  %966 = lshr <8 x i16> %965, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %967 = add <8 x i16> %958, %966
>  %968 = and <8 x i16> %967, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %969 = sub <8 x i16> %963, %959
>  %970 = mul <8 x i16> %955, %969
>  %971 = lshr <8 x i16> %970, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %972 = add <8 x i16> %959, %971
>  %973 = and <8 x i16> %972, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %974 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %968, <8 x i16> %973)
>  store <16 x i8> %974, <16 x i8>* %packed_var
>  %975 = load <16 x i8>, <16 x i8>* %packed_var
>  %976 = bitcast <16 x i8> %975 to <4 x i32>
>  %977 = and <4 x i32> %976, <i32 255, i32 255, i32 255, i32 255>
>  %978 = sitofp <4 x i32> %977 to <4 x float>
>  %979 = fmul <4 x float> %978, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %980 = lshr <4 x i32> %976, <i32 8, i32 8, i32 8, i32 8>
>  %981 = and <4 x i32> %980, <i32 255, i32 255, i32 255, i32 255>
>  %982 = sitofp <4 x i32> %981 to <4 x float>
>  %983 = fmul <4 x float> %982, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %984 = lshr <4 x i32> %976, <i32 16, i32 16, i32 16, i32 16>
>  %985 = and <4 x i32> %984, <i32 255, i32 255, i32 255, i32 255>
>  %986 = sitofp <4 x i32> %985 to <4 x float>
>  %987 = fmul <4 x float> %986, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %988 = lshr <4 x i32> %976, <i32 24, i32 24, i32 24, i32 24>
>  %989 = sitofp <4 x i32> %988 to <4 x float>
>  %990 = fmul <4 x float> %989, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %991 = shufflevector <8 x float> %728, <8 x float> %728, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %992 = shufflevector <8 x float> %729, <8 x float> %729, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  store <16 x i8> zeroinitializer, <16 x i8>* %packed_var91
>  %993 = insertelement <4 x i32> undef, i32 %context.texture0.first_level, i32 0
>  %994 = shufflevector <4 x i32> %993, <4 x i32> undef, <4 x i32> zeroinitializer
>  %minify92 = lshr <4 x i32> %734, %994
>  %995 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify92, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %996 = getelementptr [16 x i32], [16 x i32]* %context.texture0.row_stride, i32 0, i32 %context.texture0.first_level
>  %997 = load i32, i32* %996
>  %998 = insertelement <4 x i32> undef, i32 %997, i32 0
>  %999 = shufflevector <4 x i32> %998, <4 x i32> undef, <4 x i32> zeroinitializer
>  %1000 = getelementptr [16 x i32], [16 x i32]* %context.texture0.mip_offsets, i32 0, i32 %context.texture0.first_level
>  %1001 = load i32, i32* %1000
>  %1002 = getelementptr i8, i8* %context.texture0.base_ptr, i32 %1001
>  %1003 = shufflevector <4 x i32> %995, <4 x i32> undef, <4 x i32> zeroinitializer
>  %1004 = shufflevector <4 x i32> %995, <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
>  %1005 = shl <4 x i32> %995, <i32 8, i32 8, i32 8, i32 8>
>  %1006 = sitofp <4 x i32> %1005 to <4 x float>
>  %1007 = shufflevector <4 x float> %1006, <4 x float> undef, <4 x i32> zeroinitializer
>  %1008 = shufflevector <4 x float> %1006, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
>  %1009 = fmul <4 x float> %991, %1007
>  %1010 = fmul <4 x float> %992, %1008
>  %1011 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %1009)
>  %1012 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %1010)
>  %1013 = add <4 x i32> %1011, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %1014 = add <4 x i32> %1012, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %1015 = ashr <4 x i32> %1013, <i32 8, i32 8, i32 8, i32 8>
>  %1016 = ashr <4 x i32> %1014, <i32 8, i32 8, i32 8, i32 8>
>  %1017 = and <4 x i32> %1013, <i32 255, i32 255, i32 255, i32 255>
>  %1018 = and <4 x i32> %1014, <i32 255, i32 255, i32 255, i32 255>
>  %1019 = sub <4 x i32> %1003, <i32 1, i32 1, i32 1, i32 1>
>  %1020 = sitofp <4 x i32> %1003 to <4 x float>
>  %1021 = sub <4 x i32> %1003, <i32 1, i32 1, i32 1, i32 1>
>  %1022 = call <4 x float> @llvm.floor.v4f32(<4 x float> %991)
>  %1023 = fsub <4 x float> %991, %1022
>  %1024 = fmul <4 x float> %1023, %1020
>  %1025 = fmul <4 x float> %1024, <float 2.560000e+02, float 2.560000e+02, float 2.560000e+02, float 2.560000e+02>
>  %1026 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %1025)
>  %1027 = add <4 x i32> %1026, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %1028 = and <4 x i32> %1027, <i32 255, i32 255, i32 255, i32 255>
>  %1029 = ashr <4 x i32> %1027, <i32 8, i32 8, i32 8, i32 8>
>  %1030 = icmp slt <4 x i32> %1029, zeroinitializer
>  %1031 = sext <4 x i1> %1030 to <4 x i32>
>  %1032 = bitcast <4 x i32> %1031 to <16 x i8>
>  %1033 = bitcast <4 x i32> %1021 to <16 x i8>
>  %1034 = bitcast <4 x i32> %1029 to <16 x i8>
>  %1035 = call <16 x i8> @llvm.x86.sse41.pblendvb(<16 x i8> %1034, <16 x i8> %1033, <16 x i8> %1032)
>  %1036 = bitcast <16 x i8> %1035 to <4 x i32>
>  %1037 = icmp ne <4 x i32> %1036, %1019
>  %1038 = sext <4 x i1> %1037 to <4 x i32>
>  %1039 = mul <4 x i32> %1036, <i32 4, i32 4, i32 4, i32 4>
>  %1040 = add <4 x i32> %1039, <i32 4, i32 4, i32 4, i32 4>
>  %1041 = and <4 x i32> %1040, %1038
>  %1042 = sub <4 x i32> %1004, <i32 1, i32 1, i32 1, i32 1>
>  %1043 = sitofp <4 x i32> %1004 to <4 x float>
>  %1044 = sub <4 x i32> %1004, <i32 1, i32 1, i32 1, i32 1>
>  %1045 = call <4 x float> @llvm.floor.v4f32(<4 x float> %992)
>  %1046 = fsub <4 x float> %992, %1045
>  %1047 = fmul <4 x float> %1046, %1043
>  %1048 = fmul <4 x float> %1047, <float 2.560000e+02, float 2.560000e+02, float 2.560000e+02, float 2.560000e+02>
>  %1049 = call <4 x i32> @llvm.x86.sse2.cvtps2dq(<4 x float> %1048)
>  %1050 = add <4 x i32> %1049, <i32 -128, i32 -128, i32 -128, i32 -128>
>  %1051 = and <4 x i32> %1050, <i32 255, i32 255, i32 255, i32 255>
>  %1052 = ashr <4 x i32> %1050, <i32 8, i32 8, i32 8, i32 8>
>  %1053 = icmp slt <4 x i32> %1052, zeroinitializer
>  %1054 = sext <4 x i1> %1053 to <4 x i32>
>  %1055 = bitcast <4 x i32> %1054 to <16 x i8>
>  %1056 = bitcast <4 x i32> %1044 to <16 x i8>
>  %1057 = bitcast <4 x i32> %1052 to <16 x i8>
>  %1058 = call <16 x i8> @llvm.x86.sse41.pblendvb(<16 x i8> %1057, <16 x i8> %1056, <16 x i8> %1055)
>  %1059 = bitcast <16 x i8> %1058 to <4 x i32>
>  %1060 = icmp ne <4 x i32> %1059, %1042
>  %1061 = sext <4 x i1> %1060 to <4 x i32>
>  %1062 = mul <4 x i32> %1059, %999
>  %1063 = add <4 x i32> %1062, %999
>  %1064 = and <4 x i32> %1063, %1061
>  %1065 = add <4 x i32> %1039, %1062
>  %1066 = add <4 x i32> %1039, %1064
>  %1067 = add <4 x i32> %1041, %1062
>  %1068 = add <4 x i32> %1041, %1064
>  %1069 = add <4 x i32> %1039, %1062
>  %1070 = add <4 x i32> %1039, %1064
>  %1071 = add <4 x i32> %1041, %1062
>  %1072 = add <4 x i32> %1041, %1064
>  %1073 = bitcast <4 x i32> %1028 to <16 x i8>
>  %1074 = bitcast <4 x i32> %1051 to <16 x i8>
>  %1075 = shufflevector <16 x i8> %1073, <16 x i8> undef, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12>
>  %1076 = shufflevector <16 x i8> %1074, <16 x i8> undef, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4, i32 8, i32 8, i32 8, i32 8, i32 12, i32 12, i32 12, i32 12>
>  %1077 = extractelement <4 x i32> %1065, i32 0
>  %1078 = getelementptr i8, i8* %1002, i32 %1077
>  %1079 = bitcast i8* %1078 to i32*
>  %1080 = load i32, i32* %1079
>  %1081 = insertelement <4 x i32> undef, i32 %1080, i32 0
>  %1082 = extractelement <4 x i32> %1065, i32 1
>  %1083 = getelementptr i8, i8* %1002, i32 %1082
>  %1084 = bitcast i8* %1083 to i32*
>  %1085 = load i32, i32* %1084
>  %1086 = insertelement <4 x i32> %1081, i32 %1085, i32 1
>  %1087 = extractelement <4 x i32> %1065, i32 2
>  %1088 = getelementptr i8, i8* %1002, i32 %1087
>  %1089 = bitcast i8* %1088 to i32*
>  %1090 = load i32, i32* %1089
>  %1091 = insertelement <4 x i32> %1086, i32 %1090, i32 2
>  %1092 = extractelement <4 x i32> %1065, i32 3
>  %1093 = getelementptr i8, i8* %1002, i32 %1092
>  %1094 = bitcast i8* %1093 to i32*
>  %1095 = load i32, i32* %1094
>  %1096 = insertelement <4 x i32> %1091, i32 %1095, i32 3
>  %1097 = bitcast <4 x i32> %1096 to <16 x i8>
>  %1098 = extractelement <4 x i32> %1067, i32 0
>  %1099 = getelementptr i8, i8* %1002, i32 %1098
>  %1100 = bitcast i8* %1099 to i32*
>  %1101 = load i32, i32* %1100
>  %1102 = insertelement <4 x i32> undef, i32 %1101, i32 0
>  %1103 = extractelement <4 x i32> %1067, i32 1
>  %1104 = getelementptr i8, i8* %1002, i32 %1103
>  %1105 = bitcast i8* %1104 to i32*
>  %1106 = load i32, i32* %1105
>  %1107 = insertelement <4 x i32> %1102, i32 %1106, i32 1
>  %1108 = extractelement <4 x i32> %1067, i32 2
>  %1109 = getelementptr i8, i8* %1002, i32 %1108
>  %1110 = bitcast i8* %1109 to i32*
>  %1111 = load i32, i32* %1110
>  %1112 = insertelement <4 x i32> %1107, i32 %1111, i32 2
>  %1113 = extractelement <4 x i32> %1067, i32 3
>  %1114 = getelementptr i8, i8* %1002, i32 %1113
>  %1115 = bitcast i8* %1114 to i32*
>  %1116 = load i32, i32* %1115
>  %1117 = insertelement <4 x i32> %1112, i32 %1116, i32 3
>  %1118 = bitcast <4 x i32> %1117 to <16 x i8>
>  %1119 = extractelement <4 x i32> %1066, i32 0
>  %1120 = getelementptr i8, i8* %1002, i32 %1119
>  %1121 = bitcast i8* %1120 to i32*
>  %1122 = load i32, i32* %1121
>  %1123 = insertelement <4 x i32> undef, i32 %1122, i32 0
>  %1124 = extractelement <4 x i32> %1066, i32 1
>  %1125 = getelementptr i8, i8* %1002, i32 %1124
>  %1126 = bitcast i8* %1125 to i32*
>  %1127 = load i32, i32* %1126
>  %1128 = insertelement <4 x i32> %1123, i32 %1127, i32 1
>  %1129 = extractelement <4 x i32> %1066, i32 2
>  %1130 = getelementptr i8, i8* %1002, i32 %1129
>  %1131 = bitcast i8* %1130 to i32*
>  %1132 = load i32, i32* %1131
>  %1133 = insertelement <4 x i32> %1128, i32 %1132, i32 2
>  %1134 = extractelement <4 x i32> %1066, i32 3
>  %1135 = getelementptr i8, i8* %1002, i32 %1134
>  %1136 = bitcast i8* %1135 to i32*
>  %1137 = load i32, i32* %1136
>  %1138 = insertelement <4 x i32> %1133, i32 %1137, i32 3
>  %1139 = bitcast <4 x i32> %1138 to <16 x i8>
>  %1140 = extractelement <4 x i32> %1068, i32 0
>  %1141 = getelementptr i8, i8* %1002, i32 %1140
>  %1142 = bitcast i8* %1141 to i32*
>  %1143 = load i32, i32* %1142
>  %1144 = insertelement <4 x i32> undef, i32 %1143, i32 0
>  %1145 = extractelement <4 x i32> %1068, i32 1
>  %1146 = getelementptr i8, i8* %1002, i32 %1145
>  %1147 = bitcast i8* %1146 to i32*
>  %1148 = load i32, i32* %1147
>  %1149 = insertelement <4 x i32> %1144, i32 %1148, i32 1
>  %1150 = extractelement <4 x i32> %1068, i32 2
>  %1151 = getelementptr i8, i8* %1002, i32 %1150
>  %1152 = bitcast i8* %1151 to i32*
>  %1153 = load i32, i32* %1152
>  %1154 = insertelement <4 x i32> %1149, i32 %1153, i32 2
>  %1155 = extractelement <4 x i32> %1068, i32 3
>  %1156 = getelementptr i8, i8* %1002, i32 %1155
>  %1157 = bitcast i8* %1156 to i32*
>  %1158 = load i32, i32* %1157
>  %1159 = insertelement <4 x i32> %1154, i32 %1158, i32 3
>  %1160 = bitcast <4 x i32> %1159 to <16 x i8>
>  %1161 = shufflevector <16 x i8> %1075, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1162 = shufflevector <16 x i8> %1075, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1163 = bitcast <16 x i8> %1161 to <8 x i16>
>  %1164 = bitcast <16 x i8> %1162 to <8 x i16>
>  %1165 = shufflevector <16 x i8> %1097, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1166 = shufflevector <16 x i8> %1097, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1167 = bitcast <16 x i8> %1165 to <8 x i16>
>  %1168 = bitcast <16 x i8> %1166 to <8 x i16>
>  %1169 = shufflevector <16 x i8> %1118, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1170 = shufflevector <16 x i8> %1118, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1171 = bitcast <16 x i8> %1169 to <8 x i16>
>  %1172 = bitcast <16 x i8> %1170 to <8 x i16>
>  %1173 = sub <8 x i16> %1171, %1167
>  %1174 = mul <8 x i16> %1163, %1173
>  %1175 = lshr <8 x i16> %1174, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %1176 = add <8 x i16> %1167, %1175
>  %1177 = and <8 x i16> %1176, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %1178 = sub <8 x i16> %1172, %1168
>  %1179 = mul <8 x i16> %1164, %1178
>  %1180 = lshr <8 x i16> %1179, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %1181 = add <8 x i16> %1168, %1180
>  %1182 = and <8 x i16> %1181, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %1183 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %1177, <8 x i16> %1182)
>  %1184 = shufflevector <16 x i8> %1075, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1185 = shufflevector <16 x i8> %1075, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1186 = bitcast <16 x i8> %1184 to <8 x i16>
>  %1187 = bitcast <16 x i8> %1185 to <8 x i16>
>  %1188 = shufflevector <16 x i8> %1139, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1189 = shufflevector <16 x i8> %1139, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1190 = bitcast <16 x i8> %1188 to <8 x i16>
>  %1191 = bitcast <16 x i8> %1189 to <8 x i16>
>  %1192 = shufflevector <16 x i8> %1160, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1193 = shufflevector <16 x i8> %1160, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1194 = bitcast <16 x i8> %1192 to <8 x i16>
>  %1195 = bitcast <16 x i8> %1193 to <8 x i16>
>  %1196 = sub <8 x i16> %1194, %1190
>  %1197 = mul <8 x i16> %1186, %1196
>  %1198 = lshr <8 x i16> %1197, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %1199 = add <8 x i16> %1190, %1198
>  %1200 = and <8 x i16> %1199, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %1201 = sub <8 x i16> %1195, %1191
>  %1202 = mul <8 x i16> %1187, %1201
>  %1203 = lshr <8 x i16> %1202, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %1204 = add <8 x i16> %1191, %1203
>  %1205 = and <8 x i16> %1204, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %1206 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %1200, <8 x i16> %1205)
>  %1207 = shufflevector <16 x i8> %1076, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1208 = shufflevector <16 x i8> %1076, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1209 = bitcast <16 x i8> %1207 to <8 x i16>
>  %1210 = bitcast <16 x i8> %1208 to <8 x i16>
>  %1211 = shufflevector <16 x i8> %1183, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1212 = shufflevector <16 x i8> %1183, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1213 = bitcast <16 x i8> %1211 to <8 x i16>
>  %1214 = bitcast <16 x i8> %1212 to <8 x i16>
>  %1215 = shufflevector <16 x i8> %1206, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
>  %1216 = shufflevector <16 x i8> %1206, <16 x i8> zeroinitializer, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
>  %1217 = bitcast <16 x i8> %1215 to <8 x i16>
>  %1218 = bitcast <16 x i8> %1216 to <8 x i16>
>  %1219 = sub <8 x i16> %1217, %1213
>  %1220 = mul <8 x i16> %1209, %1219
>  %1221 = lshr <8 x i16> %1220, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %1222 = add <8 x i16> %1213, %1221
>  %1223 = and <8 x i16> %1222, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %1224 = sub <8 x i16> %1218, %1214
>  %1225 = mul <8 x i16> %1210, %1224
>  %1226 = lshr <8 x i16> %1225, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
>  %1227 = add <8 x i16> %1214, %1226
>  %1228 = and <8 x i16> %1227, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
>  %1229 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %1223, <8 x i16> %1228)
>  store <16 x i8> %1229, <16 x i8>* %packed_var91
>  %1230 = load <16 x i8>, <16 x i8>* %packed_var91
>  %1231 = bitcast <16 x i8> %1230 to <4 x i32>
>  %1232 = and <4 x i32> %1231, <i32 255, i32 255, i32 255, i32 255>
>  %1233 = sitofp <4 x i32> %1232 to <4 x float>
>  %1234 = fmul <4 x float> %1233, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %1235 = lshr <4 x i32> %1231, <i32 8, i32 8, i32 8, i32 8>
>  %1236 = and <4 x i32> %1235, <i32 255, i32 255, i32 255, i32 255>
>  %1237 = sitofp <4 x i32> %1236 to <4 x float>
>  %1238 = fmul <4 x float> %1237, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %1239 = lshr <4 x i32> %1231, <i32 16, i32 16, i32 16, i32 16>
>  %1240 = and <4 x i32> %1239, <i32 255, i32 255, i32 255, i32 255>
>  %1241 = sitofp <4 x i32> %1240 to <4 x float>
>  %1242 = fmul <4 x float> %1241, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %1243 = lshr <4 x i32> %1231, <i32 24, i32 24, i32 24, i32 24>
>  %1244 = sitofp <4 x i32> %1243 to <4 x float>
>  %1245 = fmul <4 x float> %1244, <float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000, float 0x3F70101020000000>
>  %1246 = shufflevector <4 x float> %979, <4 x float> %1234, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %1247 = shufflevector <4 x float> %983, <4 x float> %1238, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %1248 = shufflevector <4 x float> %987, <4 x float> %1242, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %1249 = shufflevector <4 x float> %990, <4 x float> %1245, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %1250 = load <8 x float>, <8 x float>* %temp66
>  %1251 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1252 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1250, <8 x float> %1246, <8 x float> %1251)
>  store <8 x float> %1252, <8 x float>* %temp66
>  %1253 = load <8 x float>, <8 x float>* %temp67
>  %1254 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1255 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1253, <8 x float> %1247, <8 x float> %1254)
>  store <8 x float> %1255, <8 x float>* %temp67
>  %1256 = load <8 x float>, <8 x float>* %temp68
>  %1257 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1258 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1256, <8 x float> %1248, <8 x float> %1257)
>  store <8 x float> %1258, <8 x float>* %temp68
>  %1259 = load <8 x float>, <8 x float>* %temp69
>  %1260 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1261 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1259, <8 x float> %1249, <8 x float> %1260)
>  store <8 x float> %1261, <8 x float>* %temp69
>  %1262 = load <8 x float>, <8 x float>* %temp62
>  %1263 = load <8 x float>, <8 x float>* %temp66
>  %1264 = fmul <8 x float> %1262, %1263
>  %1265 = load <8 x float>, <8 x float>* %temp63
>  %1266 = load <8 x float>, <8 x float>* %temp67
>  %1267 = fmul <8 x float> %1265, %1266
>  %1268 = load <8 x float>, <8 x float>* %temp64
>  %1269 = load <8 x float>, <8 x float>* %temp68
>  %1270 = fmul <8 x float> %1268, %1269
>  %1271 = load <8 x float>, <8 x float>* %temp65
>  %1272 = load <8 x float>, <8 x float>* %temp69
>  %1273 = fmul <8 x float> %1271, %1272
>  %1274 = load <8 x float>, <8 x float>* %temp62
>  %1275 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1276 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1274, <8 x float> %1264, <8 x float> %1275)
>  store <8 x float> %1276, <8 x float>* %temp62
>  %1277 = load <8 x float>, <8 x float>* %temp63
>  %1278 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1279 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1277, <8 x float> %1267, <8 x float> %1278)
>  store <8 x float> %1279, <8 x float>* %temp63
>  %1280 = load <8 x float>, <8 x float>* %temp64
>  %1281 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1282 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1280, <8 x float> %1270, <8 x float> %1281)
>  store <8 x float> %1282, <8 x float>* %temp64
>  %1283 = load <8 x float>, <8 x float>* %temp65
>  %1284 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1285 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1283, <8 x float> %1273, <8 x float> %1284)
>  store <8 x float> %1285, <8 x float>* %temp65
>  %1286 = load <8 x float>, <8 x float>* %temp65
>  %1287 = fcmp oge <8 x float> zeroinitializer, %1286
>  %1288 = sext <8 x i1> %1287 to <8 x i32>
>  %1289 = bitcast <8 x i32> %1288 to <8 x float>
>  %1290 = load <8 x float>, <8 x float>* %temp54
>  %1291 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1292 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1290, <8 x float> %1289, <8 x float> %1291)
>  store <8 x float> %1292, <8 x float>* %temp54
>  %1293 = load <8 x float>, <8 x float>* %temp54
>  %1294 = bitcast <8 x float> %1293 to <8 x i32>
>  %1295 = and <8 x i32> %1294, <i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216>
>  %1296 = bitcast <8 x i32> %1295 to <8 x float>
>  %1297 = load <8 x float>, <8 x float>* %temp54
>  %1298 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1299 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1297, <8 x float> %1296, <8 x float> %1298)
>  store <8 x float> %1299, <8 x float>* %temp54
>  %1300 = load <8 x float>, <8 x float>* %temp54
>  %1301 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %1300
>  %1302 = load <8 x float>, <8 x float>* %temp54
>  %1303 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %1302
>  %1304 = load <8 x float>, <8 x float>* %temp54
>  %1305 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %1304
>  %1306 = load <8 x float>, <8 x float>* %temp54
>  %1307 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %1306
>  %1308 = load <8 x float>, <8 x float>* %temp54
>  %1309 = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %1308
>  %1310 = fcmp uge <8 x float> %1309, zeroinitializer
>  %1311 = sext <8 x i1> %1310 to <8 x i32>
>  %kilp = xor <8 x i32> %callmask89, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>
>  %1312 = or <8 x i32> %1311, %kilp
>  %1313 = load <8 x i32>, <8 x i32>* %execution_mask
>  %1314 = and <8 x i32> %1313, %1312
>  store <8 x i32> %1314, <8 x i32>* %execution_mask
>  %1315 = load <8 x float>, <8 x float>* %temp62
>  %1316 = load <8 x float>, <8 x float>* %temp63
>  %1317 = load <8 x float>, <8 x float>* %temp64
>  %1318 = load <8 x float>, <8 x float>* %temp65
>  %1319 = load <8 x float>, <8 x float>* %output
>  %1320 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1321 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1319, <8 x float> %1315, <8 x float> %1320)
>  store <8 x float> %1321, <8 x float>* %output
>  %1322 = load <8 x float>, <8 x float>* %output36
>  %1323 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1324 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1322, <8 x float> %1316, <8 x float> %1323)
>  store <8 x float> %1324, <8 x float>* %output36
>  %1325 = load <8 x float>, <8 x float>* %output37
>  %1326 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1327 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1325, <8 x float> %1317, <8 x float> %1326)
>  store <8 x float> %1327, <8 x float>* %output37
>  %1328 = load <8 x float>, <8 x float>* %output38
>  %1329 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1330 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1328, <8 x float> %1318, <8 x float> %1329)
>  store <8 x float> %1330, <8 x float>* %output38
>  %1331 = load <8 x float>, <8 x float>* %temp58
>  %1332 = load <8 x float>, <8 x float>* %temp59
>  %1333 = load <8 x float>, <8 x float>* %temp60
>  %1334 = load <8 x float>, <8 x float>* %temp61
>  %1335 = load <8 x float>, <8 x float>* %output39
>  %1336 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1337 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1335, <8 x float> %1331, <8 x float> %1336)
>  store <8 x float> %1337, <8 x float>* %output39
>  %1338 = load <8 x float>, <8 x float>* %output40
>  %1339 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1340 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1338, <8 x float> %1332, <8 x float> %1339)
>  store <8 x float> %1340, <8 x float>* %output40
>  %1341 = load <8 x float>, <8 x float>* %output41
>  %1342 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1343 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1341, <8 x float> %1333, <8 x float> %1342)
>  store <8 x float> %1343, <8 x float>* %output41
>  %1344 = load <8 x float>, <8 x float>* %output42
>  %1345 = bitcast <8 x i32> %callmask89 to <8 x float>
>  %1346 = call <8 x float> @llvm.x86.avx.blendv.ps.256(<8 x float> %1344, <8 x float> %1334, <8 x float> %1345)
>  store <8 x float> %1346, <8 x float>* %output42
>  %1347 = load <8 x float>, <8 x float>* %output
>  %1348 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 0
>  store <8 x float> %1347, <8 x float>* %1348
>  %1349 = load <8 x float>, <8 x float>* %output36
>  %1350 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 1
>  store <8 x float> %1349, <8 x float>* %1350
>  %1351 = load <8 x float>, <8 x float>* %output37
>  %1352 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 2
>  store <8 x float> %1351, <8 x float>* %1352
>  %1353 = load <8 x float>, <8 x float>* %output38
>  %1354 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 0, i32 3
>  store <8 x float> %1353, <8 x float>* %1354
>  %1355 = load <8 x float>, <8 x float>* %output39
>  %1356 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 0
>  store <8 x float> %1355, <8 x float>* %1356
>  %1357 = load <8 x float>, <8 x float>* %output40
>  %1358 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 1
>  store <8 x float> %1357, <8 x float>* %1358
>  %1359 = load <8 x float>, <8 x float>* %output41
>  %1360 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 2
>  store <8 x float> %1359, <8 x float>* %1360
>  %1361 = load <8 x float>, <8 x float>* %output42
>  %1362 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 17, i32 1, i32 3
>  store <8 x float> %1361, <8 x float>* %1362
>  br label %skip
>
>skip:                                             ; preds = %entry
>  %1363 = load <8 x i32>, <8 x i32>* %execution_mask
>  %1364 = getelementptr { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }, { { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, <8 x float>, <8 x i32>, <8 x float>, <8 x i32>, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, { <8 x float>, <8 x float>, <8 x float>, <8 x float> }, float*, float*, float*, float*, float*, float, float*, float*, [8 x [4 x <8 x float>]], i32, i32, i32 }* %psCtx, i32 0, i32 3
>  store <8 x i32> %1363, <8 x i32>* %1364
>  ret void
>}
>
>FS:
>     0:		pushq	%rbp
>     1:		movq	%rsp, %rbp
>     4:		pushq	%r15
>     6:		pushq	%r14
>     8:		pushq	%r13
>    10:		pushq	%r12
>    12:		pushq	%rbx
>    13:		andq	$-32, %rsp
>    17:		subq	$704, %rsp
>    24:		movq	%rsi, 24(%rsp)
>    29:		movq	%rdi, %r10
>    32:		movq	776(%rsi), %r14
>    39:		vmovaps	416(%rsi), %ymm3
>    47:		vmovaps	%ymm3, 544(%rsp)
>    56:		vmovaps	544(%rsi), %ymm2
>    64:		vmovaps	%ymm2, 512(%rsp)
>    73:		movabsq	$140164720566276, %rax
>    83:		vbroadcastss	(%rax), %ymm1
>    88:		vmovaps	%ymm1, 192(%rsp)
>    97:		vsubps	%ymm3, %ymm1, %ymm0
>   101:		vmovaps	%ymm3, %ymm5
>   105:		vmovaps	%ymm1, %ymm14
>   109:		vsubps	%ymm2, %ymm0, %ymm4
>   113:		vmovaps	%ymm4, 576(%rsp)
>   122:		vmovaps	%ymm2, %ymm3
>   126:		vbroadcastss	8(%r14), %ymm0
>   132:		vbroadcastss	24(%r14), %ymm1
>   138:		vbroadcastss	40(%r14), %ymm2
>   144:		vmulps	%ymm2, %ymm4, %ymm2
>   148:		vmulps	%ymm0, %ymm5, %ymm0
>   152:		vmulps	%ymm1, %ymm3, %ymm1
>   156:		vaddps	%ymm1, %ymm0, %ymm0
>   160:		vaddps	%ymm2, %ymm0, %ymm15
>   164:		vbroadcastss	12(%r14), %ymm0
>   170:		vbroadcastss	28(%r14), %ymm1
>   176:		vbroadcastss	44(%r14), %ymm2
>   182:		vmulps	%ymm2, %ymm4, %ymm2
>   186:		vmulps	%ymm0, %ymm5, %ymm0
>   190:		vmulps	%ymm1, %ymm3, %ymm1
>   194:		vaddps	%ymm1, %ymm0, %ymm0
>   198:		vaddps	%ymm2, %ymm0, %ymm0
>   202:		vmovaps	%ymm0, 480(%rsp)
>   211:		vbroadcastss	48(%r14), %ymm0
>   217:		vbroadcastss	64(%r14), %ymm1
>   223:		vbroadcastss	80(%r14), %ymm2
>   229:		vmulps	%ymm2, %ymm4, %ymm2
>   233:		vmulps	%ymm0, %ymm5, %ymm0
>   237:		vmulps	%ymm1, %ymm3, %ymm1
>   241:		vaddps	%ymm1, %ymm0, %ymm0
>   245:		vaddps	%ymm2, %ymm0, %ymm0
>   249:		vmovaps	%ymm0, 416(%rsp)
>   258:		movq	384(%r10), %rax
>   265:		movq	%rax, 640(%rsp)
>   273:		vbroadcastss	112(%rax), %ymm0
>   279:		vbroadcastss	116(%rax), %ymm1
>   285:		vmulps	160(%rsi), %ymm0, %ymm0
>   293:		vaddps	%ymm1, %ymm0, %ymm13
>   297:		vmovq	9056(%r10), %xmm3
>   306:		vpbroadcastd	9068(%r10), %ymm0
>   315:		vpbroadcastd	9072(%r10), %ymm1
>   324:		vpcmpgtd	%ymm1, %ymm0, %ymm8
>   328:		vpandn	%ymm0, %ymm8, %ymm2
>   332:		vpxor	%xmm6, %xmm6, %xmm6
>   336:		vpblendw	$3, %xmm2, %xmm6, %xmm0
>   342:		vpsrld	%xmm0, %xmm3, %xmm9
>   346:		vpshufd	$229, %xmm2, %xmm0
>   351:		vpblendw	$3, %xmm0, %xmm6, %xmm0
>   357:		vpsrld	%xmm0, %xmm3, %xmm10
>   361:		vpshufd	$78, %xmm2, %xmm0
>   366:		vpblendw	$3, %xmm0, %xmm6, %xmm0
>   372:		vpsrld	%xmm0, %xmm3, %xmm12
>   376:		vpshufd	$231, %xmm2, %xmm0
>   381:		vpblendw	$3, %xmm0, %xmm6, %xmm0
>   387:		vpsrld	%xmm0, %xmm3, %xmm11
>   391:		vextracti128	$1, %ymm2, %xmm7
>   397:		vpblendw	$3, %xmm7, %xmm6, %xmm0
>   403:		vpsrld	%xmm0, %xmm3, %xmm0
>   407:		vpshufd	$229, %xmm7, %xmm1
>   412:		vpblendw	$3, %xmm1, %xmm6, %xmm1
>   418:		vpsrld	%xmm1, %xmm3, %xmm5
>   422:		vpshufd	$78, %xmm7, %xmm1
>   427:		vpblendw	$3, %xmm1, %xmm6, %xmm1
>   433:		vpsrld	%xmm1, %xmm3, %xmm1
>   437:		vpshufd	$231, %xmm7, %xmm4
>   442:		vpblendw	$3, %xmm4, %xmm6, %xmm4
>   448:		vpsrld	%xmm4, %xmm3, %xmm3
>   452:		vmovd	%xmm2, %r9d
>   457:		vpextrd	$1, %xmm2, %eax
>   463:		vpextrd	$2, %xmm2, %r11d
>   469:		vpextrd	$3, %xmm2, %r15d
>   475:		vmovd	%xmm7, %ecx
>   479:		vpextrd	$1, %xmm7, %r12d
>   485:		vpextrd	$2, %xmm7, %r13d
>   491:		vpextrd	$3, %xmm7, %r8d
>   497:		movslq	%r9d, %rdx
>   500:		movq	%rdx, 608(%rsp)
>   508:		cltq
>   510:		movq	%rax, 320(%rsp)
>   518:		movslq	%r11d, %rsi
>   521:		movq	%rsi, 288(%rsp)
>   529:		movslq	%r15d, %rdi
>   532:		movq	%rdi, 256(%rsp)
>   540:		movslq	%ecx, %rcx
>   543:		movq	%rcx, 448(%rsp)
>   551:		movslq	%r12d, %rbx
>   554:		movq	%rbx, 352(%rsp)
>   562:		movslq	%r13d, %r12
>   565:		movslq	%r8d, %r13
>   568:		vmovd	9088(%r10,%rcx,4), %xmm2
>   578:		vpinsrd	$1, 9088(%r10,%rbx,4), %xmm2, %xmm2
>   589:		vpinsrd	$2, 9088(%r10,%r12,4), %xmm2, %xmm2
>   600:		vpinsrd	$3, 9088(%r10,%r13,4), %xmm2, %xmm2
>   611:		vmovd	9088(%r10,%rdx,4), %xmm4
>   621:		vpinsrd	$1, 9088(%r10,%rax,4), %xmm4, %xmm4
>   632:		vpinsrd	$2, 9088(%r10,%rsi,4), %xmm4, %xmm4
>   643:		vpinsrd	$3, 9088(%r10,%rdi,4), %xmm4, %xmm4
>   654:		vinserti128	$1, %xmm2, %ymm4, %ymm2
>   660:		movabsq	$140164720566280, %rbx
>   670:		vpbroadcastd	(%rbx), %xmm6
>   675:		vmovdqa	%xmm6, 240(%rsp)
>   684:		vpmaxsd	%xmm6, %xmm5, %xmm4
>   689:		vmovd	%xmm4, %r15d
>   694:		vpextrd	$1, %xmm4, %ebx
>   700:		vpmaxsd	%xmm6, %xmm1, %xmm1
>   705:		vmovd	%xmm1, %edi
>   709:		vpextrd	$1, %xmm1, %r11d
>   715:		vpmaxsd	%xmm6, %xmm3, %xmm1
>   720:		vmovd	%xmm1, %ecx
>   724:		vpextrd	$1, %xmm1, %r9d
>   730:		vpmaxsd	%xmm6, %xmm10, %xmm1
>   735:		vmovd	%xmm1, %esi
>   739:		vpextrd	$1, %xmm1, %r8d
>   745:		vpmaxsd	%xmm6, %xmm12, %xmm1
>   750:		vpmaxsd	%xmm6, %xmm0, %xmm0
>   755:		vmovd	%xmm1, %eax
>   759:		vpextrd	$1, %xmm1, %edx
>   765:		vpinsrd	$1, %r15d, %xmm0, %xmm1
>   771:		vpinsrd	$2, %edi, %xmm1, %xmm1
>   777:		vpinsrd	$3, %ecx, %xmm1, %xmm1
>   783:		vpmaxsd	%xmm6, %xmm9, %xmm3
>   788:		vpmaxsd	%xmm6, %xmm11, %xmm4
>   793:		vmovd	%xmm4, %ecx
>   797:		vpextrd	$1, %xmm4, %edi
>   803:		vpinsrd	$1, %esi, %xmm3, %xmm4
>   809:		vpinsrd	$2, %eax, %xmm4, %xmm4
>   815:		vpinsrd	$3, %ecx, %xmm4, %xmm4
>   821:		vpshufd	$229, %xmm0, %xmm0
>   826:		vpinsrd	$1, %ebx, %xmm0, %xmm0
>   832:		vpinsrd	$2, %r11d, %xmm0, %xmm0
>   838:		vpinsrd	$3, %r9d, %xmm0, %xmm0
>   844:		vpshufd	$229, %xmm3, %xmm3
>   849:		vpinsrd	$1, %r8d, %xmm3, %xmm3
>   855:		vpinsrd	$2, %edx, %xmm3, %xmm3
>   861:		vpinsrd	$3, %edi, %xmm3, %xmm3
>   867:		movabsq	$140164720566272, %rax
>   877:		vinserti128	$1, %xmm1, %ymm4, %ymm1
>   883:		movq	24(%rsp), %rsi
>   888:		vcvttps2dq	32(%rsi), %ymm6
>   893:		vmovdqa	%ymm6, 384(%rsp)
>   902:		vcvttps2dq	%ymm13, %ymm5
>   907:		vinserti128	$1, %xmm0, %ymm3, %ymm0
>   913:		vpcmpgtd	%ymm6, %ymm1, %ymm1
>   917:		vpor	%ymm6, %ymm5, %ymm3
>   921:		vpsrad	$31, %ymm3, %ymm3
>   926:		vpor	%ymm8, %ymm3, %ymm3
>   931:		vpcmpeqd	%ymm4, %ymm4, %ymm4
>   935:		vpxor	%ymm4, %ymm1, %ymm1
>   939:		vpor	%ymm1, %ymm3, %ymm1
>   943:		vmovss	(%rax), %xmm3
>   947:		vmovaps	%xmm3, 176(%rsp)
>   956:		movq	9080(%r10), %rbx
>   963:		vpcmpgtd	%ymm5, %ymm0, %ymm0
>   967:		vpxor	%ymm4, %ymm0, %ymm0
>   971:		vpor	%ymm0, %ymm1, %ymm4
>   975:		vpslld	$3, %ymm6, %ymm0
>   980:		vpmulld	%ymm5, %ymm2, %ymm1
>   985:		vpaddd	%ymm0, %ymm1, %ymm0
>   989:		movq	448(%rsp), %rax
>   997:		vmovd	9216(%r10,%rax,4), %xmm1
>  1007:		movq	352(%rsp), %rax
>  1015:		vpinsrd	$1, 9216(%r10,%rax,4), %xmm1, %xmm1
>  1026:		vpinsrd	$2, 9216(%r10,%r12,4), %xmm1, %xmm1
>  1037:		vpinsrd	$3, 9216(%r10,%r13,4), %xmm1, %xmm1
>  1048:		movq	608(%rsp), %rax
>  1056:		vmovd	9216(%r10,%rax,4), %xmm2
>  1066:		movq	320(%rsp), %rax
>  1074:		vpinsrd	$1, 9216(%r10,%rax,4), %xmm2, %xmm2
>  1085:		movq	288(%rsp), %rax
>  1093:		vpinsrd	$2, 9216(%r10,%rax,4), %xmm2, %xmm2
>  1104:		movq	256(%rsp), %rax
>  1112:		vpinsrd	$3, 9216(%r10,%rax,4), %xmm2, %xmm2
>  1123:		vinserti128	$1, %xmm1, %ymm2, %ymm1
>  1129:		vpaddd	%ymm1, %ymm0, %ymm0
>  1133:		vpandn	%ymm0, %ymm4, %ymm1
>  1137:		vmovq	%xmm1, %r9
>  1142:		movslq	%r9d, %rax
>  1145:		vmovq	(%rbx,%rax), %xmm0
>  1150:		vmovshdup	%xmm0, %xmm0
>  1154:		sarq	$32, %r9
>  1158:		vpextrq	$1, %xmm1, %rcx
>  1164:		vextracti128	$1, %ymm1, %xmm1
>  1170:		vmovq	%xmm1, %rax
>  1175:		movslq	%eax, %rdx
>  1178:		vmovq	(%rbx,%rdx), %xmm2
>  1183:		vmovshdup	%xmm2, %xmm2
>  1187:		sarq	$32, %rax
>  1191:		vpextrq	$1, %xmm1, %rdx
>  1197:		vmovq	(%rbx,%rax), %xmm1
>  1202:		vmovshdup	%xmm1, %xmm1
>  1206:		movslq	%edx, %rax
>  1209:		vinsertps	$16, %xmm1, %xmm2, %xmm1
>  1215:		vmovq	(%rbx,%rax), %xmm2
>  1220:		vmovshdup	%xmm2, %xmm2
>  1224:		sarq	$32, %rdx
>  1228:		vinsertps	$32, %xmm2, %xmm1, %xmm1
>  1234:		vmovq	(%rbx,%rdx), %xmm2
>  1239:		vmovshdup	%xmm2, %xmm2
>  1243:		vinsertps	$48, %xmm2, %xmm1, %xmm1
>  1249:		vmovq	(%rbx,%r9), %xmm2
>  1255:		vmovshdup	%xmm2, %xmm2
>  1259:		movslq	%ecx, %rax
>  1262:		vinsertps	$16, %xmm2, %xmm0, %xmm0
>  1268:		vmovq	(%rbx,%rax), %xmm2
>  1273:		vmovshdup	%xmm2, %xmm2
>  1277:		sarq	$32, %rcx
>  1281:		vinsertps	$32, %xmm2, %xmm0, %xmm0
>  1287:		vmovq	(%rbx,%rcx), %xmm2
>  1292:		vmovshdup	%xmm2, %xmm2
>  1296:		vinsertps	$48, %xmm2, %xmm0, %xmm0
>  1302:		movq	%rsi, %rcx
>  1305:		vmovaps	256(%rcx), %ymm8
>  1313:		vinsertf128	$1, %xmm1, %ymm0, %ymm0
>  1319:		vpxor	%ymm12, %ymm12, %ymm12
>  1324:		vpandn	%ymm0, %ymm4, %ymm7
>  1328:		vbroadcastss	%xmm3, %ymm5
>  1333:		vmovaps	%ymm5, 288(%rsp)
>  1342:		vcmpneqps	%ymm5, %ymm7, %ymm0
>  1347:		vcmpltps	%ymm8, %ymm7, %ymm1
>  1353:		vmovaps	%ymm1, 352(%rsp)
>  1362:		vblendvps	%ymm1, %ymm0, %ymm12, %ymm0
>  1368:		vpcmpeqd	%ymm12, %ymm0, %ymm11
>  1373:		vpcmpeqd	%ymm2, %ymm2, %ymm2
>  1377:		vpxor	%ymm2, %ymm11, %ymm4
>  1381:		vblendvps	%ymm4, %ymm5, %ymm0, %ymm10
>  1387:		movabsq	$140164720566284, %rax
>  1397:		vbroadcastss	(%rax), %ymm0
>  1402:		vcmpltps	%ymm0, %ymm8, %ymm0
>  1407:		vpxor	%ymm2, %ymm4, %ymm2
>  1411:		vblendvps	%ymm2, %ymm0, %ymm10, %ymm2
>  1417:		vpcmpeqd	%ymm12, %ymm2, %ymm0
>  1422:		vpandn	%ymm11, %ymm0, %ymm0
>  1427:		movabsq	$140164720566368, %rax
>  1437:		vmovaps	(%rax), %ymm9
>  1441:		vmovaps	%ymm9, 608(%rsp)
>  1450:		vxorps	%ymm9, %ymm8, %ymm5
>  1455:		vblendvps	%ymm0, %ymm5, %ymm2, %ymm3
>  1461:		vmovaps	%ymm3, 96(%rsp)
>  1467:		vmovaps	%ymm14, %ymm1
>  1471:		vdivps	672(%rcx), %ymm1, %ymm6
>  1479:		vmulps	%ymm15, %ymm6, %ymm13
>  1484:		vmovaps	%ymm13, 448(%rsp)
>  1493:		vmulps	480(%rsp), %ymm6, %ymm2
>  1502:		vmovaps	%ymm2, 256(%rsp)
>  1511:		vmulps	%ymm13, %ymm13, %ymm14
>  1516:		vmulps	%ymm2, %ymm2, %ymm15
>  1520:		vaddps	%ymm15, %ymm14, %ymm2
>  1525:		vmulps	416(%rsp), %ymm6, %ymm15
>  1534:		vmulps	%ymm15, %ymm15, %ymm14
>  1539:		vaddps	%ymm14, %ymm2, %ymm14
>  1544:		movl	1856(%rcx), %eax
>  1550:		movq	%rcx, %r15
>  1553:		vpandn	%ymm11, %ymm0, %ymm11
>  1558:		vcvtsi2ssq	%rax, %xmm0, %xmm5
>  1563:		vblendvps	%ymm11, %ymm14, %ymm3, %ymm14
>  1569:		vsqrtps	%ymm14, %ymm3
>  1574:		vaddss	%xmm5, %xmm5, %xmm5
>  1578:		vdivps	%ymm3, %ymm1, %ymm3
>  1582:		vaddss	176(%rsp), %xmm5, %xmm5
>  1591:		vblendvps	%ymm11, %ymm3, %ymm14, %ymm14
>  1597:		vbroadcastss	%xmm5, %ymm3
>  1602:		vcmpleps	%ymm3, %ymm12, %ymm5
>  1607:		vmulps	%ymm14, %ymm15, %ymm3
>  1612:		vblendvps	%ymm11, %ymm3, %ymm12, %ymm1
>  1618:		vmovaps	%ymm1, 320(%rsp)
>  1627:		vpcmpeqd	%ymm12, %ymm5, %ymm5
>  1632:		vblendvps	%ymm11, %ymm5, %ymm12, %ymm5
>  1638:		vpcmpeqd	%ymm12, %ymm5, %ymm5
>  1643:		vpcmpeqd	%ymm3, %ymm3, %ymm3
>  1647:		vpxor	%ymm3, %ymm5, %ymm5
>  1651:		vblendvps	%ymm5, %ymm11, %ymm12, %ymm5
>  1657:		vblendvps	%ymm11, %ymm1, %ymm12, %ymm15
>  1663:		vxorps	%ymm9, %ymm1, %ymm3
>  1668:		vblendvps	%ymm5, %ymm3, %ymm15, %ymm9
>  1674:		vmovaps	%ymm9, 416(%rsp)
>  1683:		vbroadcastss	(%r14), %ymm3
>  1688:		vmovaps	544(%rsp), %ymm1
>  1697:		vmulps	%ymm3, %ymm1, %ymm3
>  1701:		vbroadcastss	16(%r14), %ymm5
>  1707:		vmovaps	512(%rsp), %ymm2
>  1716:		vmulps	%ymm5, %ymm2, %ymm5
>  1720:		vaddps	%ymm5, %ymm3, %ymm3
>  1724:		vbroadcastss	32(%r14), %ymm5
>  1730:		vmovaps	576(%rsp), %ymm12
>  1739:		vmulps	%ymm5, %ymm12, %ymm5
>  1743:		vaddps	%ymm5, %ymm3, %ymm3
>  1747:		vbroadcastss	36(%r14), %ymm5
>  1753:		vmulps	%ymm5, %ymm12, %ymm5
>  1757:		vbroadcastss	4(%r14), %ymm15
>  1763:		vmulps	%ymm15, %ymm1, %ymm15
>  1768:		vbroadcastss	20(%r14), %ymm12
>  1774:		vmulps	%ymm12, %ymm2, %ymm12
>  1779:		vaddps	%ymm12, %ymm15, %ymm12
>  1784:		vaddps	%ymm5, %ymm12, %ymm5
>  1788:		vmulps	%ymm3, %ymm6, %ymm15
>  1792:		vmulps	%ymm5, %ymm6, %ymm1
>  1796:		vmovaps	%ymm1, 128(%rsp)
>  1805:		vmovaps	384(%rsp), %ymm1
>  1814:		vblendvps	%ymm4, %ymm10, %ymm1, %ymm3
>  1820:		vxorps	%ymm10, %ymm10, %ymm10
>  1825:		vblendvps	%ymm4, %ymm7, %ymm10, %ymm2
>  1831:		vblendvps	%ymm4, %ymm2, %ymm7, %ymm5
>  1837:		vblendvps	%ymm0, %ymm8, %ymm2, %ymm2
>  1843:		vblendvps	%ymm0, 96(%rsp), %ymm3, %ymm6
>  1851:		vmovaps	%ymm6, 480(%rsp)
>  1860:		vblendvps	%ymm4, %ymm3, %ymm10, %ymm3
>  1866:		vblendvps	%ymm4, %ymm5, %ymm10, %ymm1
>  1872:		vblendvps	%ymm0, %ymm2, %ymm5, %ymm5
>  1878:		vmovaps	%ymm5, 544(%rsp)
>  1887:		vblendvps	%ymm0, %ymm6, %ymm3, %ymm3
>  1893:		vmovaps	%ymm3, 512(%rsp)
>  1902:		vblendvps	%ymm0, %ymm5, %ymm1, %ymm0
>  1908:		vmovaps	%ymm0, 576(%rsp)
>  1917:		vmulps	256(%rsp), %ymm14, %ymm0
>  1926:		vmulps	%ymm14, %ymm13, %ymm1
>  1931:		vblendvps	%ymm11, %ymm1, %ymm14, %ymm1
>  1937:		vmovaps	%ymm9, %ymm8
>  1942:		vmaxps	%ymm8, %ymm10, %ymm3
>  1947:		vcmpunordps	%ymm10, %ymm8, %ymm5
>  1953:		vblendvps	%ymm5, %ymm10, %ymm3, %ymm3
>  1959:		vxorps	%ymm14, %ymm14, %ymm14
>  1964:		vblendvps	%ymm11, %ymm3, %ymm1, %ymm6
>  1970:		movabsq	$140164720566288, %rax
>  1980:		vbroadcastss	(%rax), %ymm1
>  1985:		vandps	%ymm1, %ymm6, %ymm1
>  1989:		vpsrld	$23, %ymm1, %ymm1
>  1994:		movabsq	$140164720566292, %rax
>  2004:		movabsq	$140164720566296, %rcx
>  2014:		vbroadcastss	(%rcx), %ymm3
>  2019:		vandps	%ymm3, %ymm6, %ymm3
>  2023:		movabsq	$140164720566276, %rcx
>  2033:		vbroadcastss	(%rcx), %ymm5
>  2038:		vmovaps	%ymm5, 384(%rsp)
>  2047:		vorps	%ymm5, %ymm3, %ymm3
>  2051:		vmovaps	%ymm5, %ymm12
>  2055:		vmovaps	192(%rsp), %ymm7
>  2064:		vaddps	%ymm7, %ymm3, %ymm5
>  2068:		vdivps	%ymm5, %ymm7, %ymm5
>  2072:		vmovaps	%ymm7, %ymm13
>  2076:		vpbroadcastd	(%rax), %ymm7
>  2081:		vpaddd	%ymm7, %ymm1, %ymm1
>  2085:		vaddps	288(%rsp), %ymm3, %ymm3
>  2094:		vmulps	%ymm3, %ymm5, %ymm3
>  2098:		movabsq	$140164720566300, %rax
>  2108:		vmulps	%ymm3, %ymm3, %ymm5
>  2112:		vmulps	%ymm5, %ymm5, %ymm7
>  2116:		vbroadcastss	(%rax), %ymm8
>  2121:		vmulps	%ymm8, %ymm7, %ymm8
>  2126:		movabsq	$140164720566304, %rax
>  2136:		vbroadcastss	(%rax), %ymm9
>  2141:		vaddps	%ymm9, %ymm8, %ymm8
>  2146:		movabsq	$140164720566308, %rax
>  2156:		vbroadcastss	(%rax), %ymm9
>  2161:		vmulps	%ymm9, %ymm7, %ymm9
>  2166:		movabsq	$140164720566312, %rax
>  2176:		vbroadcastss	(%rax), %ymm10
>  2181:		vaddps	%ymm10, %ymm9, %ymm9
>  2186:		vmulps	%ymm8, %ymm7, %ymm7
>  2191:		movabsq	$140164720566316, %rax
>  2201:		vbroadcastss	(%rax), %ymm8
>  2206:		vaddps	%ymm8, %ymm7, %ymm7
>  2211:		vmulps	%ymm9, %ymm5, %ymm5
>  2216:		vaddps	%ymm7, %ymm5, %ymm5
>  2220:		vmulps	%ymm5, %ymm3, %ymm3
>  2224:		vcvtdq2ps	%ymm1, %ymm1
>  2228:		vaddps	%ymm3, %ymm1, %ymm1
>  2232:		movq	640(%rsp), %rcx
>  2240:		vbroadcastss	64(%rcx), %ymm3
>  2246:		vmulps	%ymm1, %ymm3, %ymm1
>  2250:		movabsq	$140164720566320, %rax
>  2260:		vbroadcastss	(%rax), %ymm3
>  2265:		vminps	%ymm1, %ymm3, %ymm1
>  2269:		movabsq	$140164720566324, %rax
>  2279:		vbroadcastss	(%rax), %ymm3
>  2284:		vmaxps	%ymm1, %ymm3, %ymm3
>  2288:		vroundps	$1, %ymm3, %ymm5
>  2294:		vblendvps	%ymm11, %ymm0, %ymm2, %ymm1
>  2300:		vsubps	%ymm5, %ymm3, %ymm0
>  2304:		movabsq	$140164720566328, %rax
>  2314:		vmulps	%ymm0, %ymm0, %ymm2
>  2318:		vbroadcastss	(%rax), %ymm3
>  2323:		vmulps	%ymm3, %ymm2, %ymm3
>  2327:		movabsq	$140164720566332, %rax
>  2337:		vbroadcastss	(%rax), %ymm7
>  2342:		vaddps	%ymm7, %ymm3, %ymm3
>  2346:		movabsq	$140164720566336, %rax
>  2356:		vbroadcastss	(%rax), %ymm7
>  2361:		vmulps	%ymm7, %ymm2, %ymm7
>  2365:		movabsq	$140164720566340, %rax
>  2375:		vbroadcastss	(%rax), %ymm8
>  2380:		vaddps	%ymm8, %ymm7, %ymm7
>  2385:		vmulps	%ymm3, %ymm2, %ymm3
>  2389:		movabsq	$140164720566344, %rax
>  2399:		vbroadcastss	(%rax), %ymm8
>  2404:		vaddps	%ymm8, %ymm3, %ymm3
>  2409:		vmulps	%ymm7, %ymm2, %ymm2
>  2413:		vmulps	%ymm3, %ymm0, %ymm0
>  2417:		vaddps	%ymm13, %ymm2, %ymm2
>  2422:		vaddps	%ymm0, %ymm2, %ymm0
>  2426:		vcvttps2dq	%ymm5, %ymm2
>  2430:		vpslld	$23, %ymm2, %ymm2
>  2435:		vpaddd	%ymm12, %ymm2, %ymm2
>  2440:		vmulps	%ymm0, %ymm2, %ymm0
>  2444:		vmovaps	352(%rsp), %ymm2
>  2453:		vblendvps	%ymm11, %ymm0, %ymm2, %ymm0
>  2459:		movq	%rcx, %rax
>  2462:		vbroadcastss	32(%rax), %ymm2
>  2468:		vmulps	%ymm2, %ymm6, %ymm2
>  2472:		vbroadcastss	16(%rax), %ymm3
>  2478:		vaddps	%ymm3, %ymm2, %ymm2
>  2482:		vbroadcastss	36(%rax), %ymm3
>  2488:		vmulps	%ymm3, %ymm6, %ymm3
>  2492:		vbroadcastss	20(%rax), %ymm5
>  2498:		vaddps	%ymm5, %ymm3, %ymm3
>  2502:		vbroadcastss	40(%rax), %ymm5
>  2508:		vmulps	%ymm5, %ymm6, %ymm5
>  2512:		vbroadcastss	24(%rax), %ymm7
>  2518:		vaddps	%ymm7, %ymm5, %ymm5
>  2522:		vblendvps	%ymm11, %ymm2, %ymm6, %ymm2
>  2528:		vblendvps	%ymm11, %ymm3, %ymm1, %ymm1
>  2534:		vmovaps	320(%rsp), %ymm3
>  2543:		vblendvps	%ymm11, %ymm5, %ymm3, %ymm3
>  2549:		vbroadcastss	48(%rax), %ymm5
>  2555:		vmulps	%ymm5, %ymm0, %ymm5
>  2559:		vaddps	%ymm5, %ymm2, %ymm5
>  2563:		vblendvps	%ymm11, %ymm5, %ymm2, %ymm2
>  2569:		vbroadcastss	52(%rax), %ymm5
>  2575:		vmulps	%ymm5, %ymm0, %ymm5
>  2579:		vaddps	%ymm5, %ymm1, %ymm5
>  2583:		vblendvps	%ymm11, %ymm5, %ymm1, %ymm1
>  2589:		vbroadcastss	56(%rax), %ymm5
>  2595:		vmulps	%ymm5, %ymm0, %ymm5
>  2599:		vaddps	%ymm5, %ymm3, %ymm5
>  2603:		vblendvps	%ymm11, %ymm5, %ymm3, %ymm3
>  2609:		vmaxps	%ymm14, %ymm2, %ymm5
>  2614:		vmovaps	%ymm13, %ymm6
>  2618:		vminps	%ymm6, %ymm5, %ymm5
>  2622:		vblendvps	%ymm11, %ymm5, %ymm2, %ymm2
>  2628:		vmovaps	%ymm2, 320(%rsp)
>  2637:		vmaxps	%ymm14, %ymm1, %ymm2
>  2642:		vminps	%ymm6, %ymm2, %ymm2
>  2646:		vblendvps	%ymm11, %ymm2, %ymm1, %ymm1
>  2652:		vmovaps	%ymm1, 352(%rsp)
>  2661:		vmaxps	%ymm14, %ymm3, %ymm1
>  2666:		vminps	%ymm6, %ymm1, %ymm1
>  2670:		vblendvps	%ymm11, %ymm1, %ymm3, %ymm1
>  2676:		vmovaps	%ymm1, 288(%rsp)
>  2685:		vbroadcastss	(%rax), %ymm1
>  2690:		vblendvps	%ymm11, %ymm1, %ymm14, %ymm1
>  2696:		vmaxps	%ymm14, %ymm1, %ymm2
>  2701:		vminps	%ymm6, %ymm2, %ymm2
>  2705:		vblendvps	%ymm11, %ymm2, %ymm1, %ymm1
>  2711:		vmovaps	%ymm1, 256(%rsp)
>  2720:		vblendvps	%ymm11, %ymm15, %ymm0, %ymm3
>  2726:		vmovaps	%ymm3, 640(%rsp)
>  2735:		movslq	8844(%r10), %rax
>  2742:		vmovq	8832(%r10), %xmm0
>  2751:		vmovd	%eax, %xmm1
>  2755:		vpsrld	%xmm1, %xmm0, %xmm0
>  2759:		vpmaxsd	240(%rsp), %xmm0, %xmm1
>  2769:		vpbroadcastd	%xmm1, %xmm0
>  2774:		vpcmpeqd	%xmm8, %xmm8, %xmm8
>  2779:		vpaddd	%xmm8, %xmm0, %xmm6
>  2784:		vmovdqa	%xmm6, 192(%rsp)
>  2793:		vcvtdq2ps	%xmm0, %xmm5
>  2797:		vmovaps	%xmm5, 240(%rsp)
>  2806:		vroundps	$1, %xmm3, %xmm0
>  2812:		vsubps	%xmm0, %xmm3, %xmm0
>  2816:		vmulps	%xmm0, %xmm5, %xmm0
>  2820:		movabsq	$140164720566348, %rcx
>  2830:		vbroadcastss	(%rcx), %xmm3
>  2835:		vmovaps	%xmm3, 48(%rsp)
>  2841:		vmulps	%xmm3, %xmm0, %xmm0
>  2845:		vmovaps	%xmm3, %xmm4
>  2849:		vcvtps2dq	%xmm0, %xmm0
>  2853:		movabsq	$140164720566352, %rcx
>  2863:		vpbroadcastd	(%rcx), %xmm3
>  2868:		vmovdqa	%xmm3, 32(%rsp)
>  2874:		vpaddd	%xmm3, %xmm0, %xmm0
>  2878:		vmovdqa	%xmm3, %xmm2
>  2882:		vpsrad	$8, %xmm0, %xmm3
>  2887:		vpsrad	$31, %xmm0, %xmm5
>  2892:		vpblendvb	%xmm5, %xmm6, %xmm3, %xmm3
>  2898:		vpcmpeqd	%xmm6, %xmm3, %xmm5
>  2902:		vpslld	$2, %xmm3, %xmm3
>  2907:		movabsq	$140164720566360, %rcx
>  2917:		vpbroadcastd	(%rcx), %xmm6
>  2922:		vmovdqa	%xmm6, 176(%rsp)
>  2931:		vpaddd	%xmm6, %xmm3, %xmm6
>  2935:		vpandn	%xmm6, %xmm5, %xmm9
>  2939:		vblendvps	%ymm11, 128(%rsp), %ymm14, %ymm5
>  2950:		vmovaps	%ymm5, 64(%rsp)
>  2956:		vpshufd	$85, %xmm1, %xmm1
>  2961:		vpaddd	%xmm8, %xmm1, %xmm6
>  2966:		vmovdqa	%xmm6, 96(%rsp)
>  2972:		vcvtdq2ps	%xmm1, %xmm7
>  2976:		vmovaps	%xmm7, 128(%rsp)
>  2985:		vroundps	$1, %xmm5, %xmm1
>  2991:		vsubps	%xmm1, %xmm5, %xmm1
>  2995:		vmulps	%xmm1, %xmm7, %xmm1
>  2999:		vmulps	%xmm4, %xmm1, %xmm1
>  3003:		vcvtps2dq	%xmm1, %xmm1
>  3007:		vpaddd	%xmm2, %xmm1, %xmm4
>  3011:		vpsrad	$8, %xmm4, %xmm1
>  3016:		vpsrad	$31, %xmm4, %xmm2
>  3021:		vpblendvb	%xmm2, %xmm6, %xmm1, %xmm1
>  3027:		vpbroadcastd	8864(%r10,%rax,4), %xmm15
>  3037:		vpcmpeqd	%xmm6, %xmm1, %xmm2
>  3041:		vpmulld	%xmm15, %xmm1, %xmm1
>  3046:		vpaddd	%xmm15, %xmm1, %xmm6
>  3051:		vpandn	%xmm6, %xmm2, %xmm2
>  3055:		vpaddd	%xmm3, %xmm1, %xmm6
>  3059:		vpextrq	$1, %xmm6, %rdi
>  3065:		vmovq	%xmm6, %rbx
>  3070:		vpaddd	%xmm3, %xmm2, %xmm3
>  3074:		vpextrq	$1, %xmm3, %r8
>  3080:		vmovq	%xmm3, %r11
>  3085:		vpaddd	%xmm9, %xmm1, %xmm1
>  3090:		vpextrq	$1, %xmm1, %rsi
>  3096:		vmovq	%xmm1, %rdx
>  3101:		vpaddd	%xmm9, %xmm2, %xmm1
>  3106:		vpextrq	$1, %xmm1, %r9
>  3112:		vmovq	%xmm1, %r14
>  3117:		movslq	8992(%r10,%rax,4), %rax
>  3125:		addq	8856(%r10), %rax
>  3132:		movabsq	$140164720566356, %rcx
>  3142:		vpbroadcastd	(%rcx), %xmm9
>  3147:		vpand	%xmm9, %xmm0, %xmm0
>  3152:		movabsq	$140164720566400, %rcx
>  3162:		vmovdqa	(%rcx), %xmm12
>  3166:		vpshufb	%xmm12, %xmm0, %xmm3
>  3171:		movslq	%ebx, %r10
>  3174:		sarq	$32, %rbx
>  3178:		movslq	%edi, %rcx
>  3181:		sarq	$32, %rdi
>  3185:		vmovd	(%r10,%rax), %xmm0
>  3191:		vpinsrd	$1, (%rbx,%rax), %xmm0, %xmm0
>  3198:		vpinsrd	$2, (%rcx,%rax), %xmm0, %xmm0
>  3205:		vpinsrd	$3, (%rdi,%rax), %xmm0, %xmm2
>  3212:		movslq	%edx, %rcx
>  3215:		sarq	$32, %rdx
>  3219:		movslq	%esi, %rdi
>  3222:		sarq	$32, %rsi
>  3226:		vmovd	(%rcx,%rax), %xmm0
>  3231:		vpinsrd	$1, (%rdx,%rax), %xmm0, %xmm0
>  3238:		vpinsrd	$2, (%rdi,%rax), %xmm0, %xmm0
>  3245:		vpinsrd	$3, (%rsi,%rax), %xmm0, %xmm1
>  3252:		vpmovzxbw	%xmm2, %xmm6
>  3257:		vpmovzxbw	%xmm1, %xmm14
>  3262:		vpsubw	%xmm6, %xmm14, %xmm14
>  3266:		vpmovzxbw	%xmm3, %xmm0
>  3271:		vpmullw	%xmm0, %xmm14, %xmm7
>  3275:		vpsrlw	$8, %xmm7, %xmm7
>  3280:		vpaddw	%xmm6, %xmm7, %xmm6
>  3284:		vxorps	%xmm5, %xmm5, %xmm5
>  3288:		vpunpckhbw	%xmm5, %xmm3, %xmm7
>  3292:		vpunpckhbw	%xmm5, %xmm2, %xmm2
>  3296:		vpunpckhbw	%xmm5, %xmm1, %xmm1
>  3300:		vxorps	%xmm10, %xmm10, %xmm10
>  3305:		vpsubw	%xmm2, %xmm1, %xmm1
>  3309:		vpmullw	%xmm7, %xmm1, %xmm1
>  3313:		vpsrlw	$8, %xmm1, %xmm1
>  3318:		vpaddw	%xmm2, %xmm1, %xmm1
>  3322:		movabsq	$140164720566416, %rcx
>  3332:		vmovdqa	(%rcx), %xmm8
>  3336:		vpand	%xmm8, %xmm6, %xmm2
>  3341:		vpand	%xmm8, %xmm1, %xmm1
>  3346:		vpackuswb	%xmm1, %xmm2, %xmm14
>  3350:		movslq	%r11d, %rcx
>  3353:		sarq	$32, %r11
>  3357:		movslq	%r8d, %rdx
>  3360:		sarq	$32, %r8
>  3364:		vmovd	(%rcx,%rax), %xmm2
>  3369:		vpinsrd	$1, (%r11,%rax), %xmm2, %xmm2
>  3376:		vpinsrd	$2, (%rdx,%rax), %xmm2, %xmm2
>  3383:		vpinsrd	$3, (%r8,%rax), %xmm2, %xmm2
>  3390:		movslq	%r14d, %rcx
>  3393:		sarq	$32, %r14
>  3397:		movslq	%r9d, %rdx
>  3400:		sarq	$32, %r9
>  3404:		vmovd	(%rcx,%rax), %xmm6
>  3409:		vpinsrd	$1, (%r14,%rax), %xmm6, %xmm6
>  3416:		vpinsrd	$2, (%rdx,%rax), %xmm6, %xmm6
>  3423:		vpinsrd	$3, (%r9,%rax), %xmm6, %xmm6
>  3430:		vpmovzxbw	%xmm2, %xmm1
>  3435:		vpmovzxbw	%xmm6, %xmm13
>  3440:		vpsubw	%xmm1, %xmm13, %xmm5
>  3444:		vpmullw	%xmm0, %xmm5, %xmm0
>  3448:		vpsrlw	$8, %xmm0, %xmm0
>  3453:		vpaddw	%xmm1, %xmm0, %xmm0
>  3457:		vpunpckhbw	%xmm10, %xmm2, %xmm1
>  3462:		vpunpckhbw	%xmm10, %xmm6, %xmm2
>  3467:		vpsubw	%xmm1, %xmm2, %xmm2
>  3471:		vpmullw	%xmm7, %xmm2, %xmm2
>  3475:		vpsrlw	$8, %xmm2, %xmm2
>  3480:		vpaddw	%xmm1, %xmm2, %xmm1
>  3484:		vpand	%xmm8, %xmm0, %xmm0
>  3489:		vpand	%xmm8, %xmm1, %xmm1
>  3494:		vpackuswb	%xmm1, %xmm0, %xmm0
>  3498:		vpand	%xmm9, %xmm4, %xmm1
>  3503:		vpshufb	%xmm12, %xmm1, %xmm1
>  3508:		vpmovzxbw	%xmm14, %xmm2
>  3513:		vpmovzxbw	%xmm0, %xmm5
>  3518:		vpsubw	%xmm2, %xmm5, %xmm5
>  3522:		vpmovzxbw	%xmm1, %xmm6
>  3527:		vpmullw	%xmm6, %xmm5, %xmm5
>  3531:		vpsrlw	$8, %xmm5, %xmm5
>  3536:		vpaddw	%xmm2, %xmm5, %xmm2
>  3540:		vpunpckhbw	%xmm10, %xmm1, %xmm1
>  3545:		vpunpckhbw	%xmm10, %xmm14, %xmm5
>  3550:		vpunpckhbw	%xmm10, %xmm0, %xmm0
>  3555:		vpsubw	%xmm5, %xmm0, %xmm0
>  3559:		vpmullw	%xmm1, %xmm0, %xmm0
>  3563:		vpsrlw	$8, %xmm0, %xmm0
>  3568:		vpaddw	%xmm5, %xmm0, %xmm0
>  3572:		vpand	%xmm8, %xmm2, %xmm1
>  3577:		vpand	%xmm8, %xmm0, %xmm0
>  3582:		vpackuswb	%xmm0, %xmm1, %xmm13
>  3586:		vmovaps	640(%rsp), %ymm0
>  3595:		vextractf128	$1, %ymm0, %xmm0
>  3601:		vroundps	$1, %xmm0, %xmm2
>  3607:		vsubps	%xmm2, %xmm0, %xmm0
>  3611:		vmulps	240(%rsp), %xmm0, %xmm0
>  3620:		vmovaps	48(%rsp), %xmm3
>  3626:		vmulps	%xmm3, %xmm0, %xmm0
>  3630:		vcvtps2dq	%xmm0, %xmm0
>  3634:		vmovdqa	32(%rsp), %xmm4
>  3640:		vpaddd	%xmm4, %xmm0, %xmm2
>  3644:		vpsrad	$8, %xmm2, %xmm0
>  3649:		vpsrad	$31, %xmm2, %xmm5
>  3654:		vmovdqa	192(%rsp), %xmm1
>  3663:		vpblendvb	%xmm5, %xmm1, %xmm0, %xmm0
>  3669:		vpcmpeqd	%xmm1, %xmm0, %xmm5
>  3673:		vpslld	$2, %xmm0, %xmm0
>  3678:		vpaddd	176(%rsp), %xmm0, %xmm6
>  3687:		vpandn	%xmm6, %xmm5, %xmm5
>  3691:		vmovaps	64(%rsp), %ymm14
>  3697:		vextractf128	$1, %ymm14, %xmm6
>  3703:		vroundps	$1, %xmm6, %xmm7
>  3709:		vsubps	%xmm7, %xmm6, %xmm6
>  3713:		vmulps	128(%rsp), %xmm6, %xmm6
>  3722:		vmulps	%xmm3, %xmm6, %xmm6
>  3726:		vcvtps2dq	%xmm6, %xmm6
>  3730:		vpaddd	%xmm4, %xmm6, %xmm6
>  3734:		vpsrad	$8, %xmm6, %xmm7
>  3739:		vpsrad	$31, %xmm6, %xmm1
>  3744:		vmovdqa	96(%rsp), %xmm3
>  3750:		vpblendvb	%xmm1, %xmm3, %xmm7, %xmm1
>  3756:		vpcmpeqd	%xmm3, %xmm1, %xmm7
>  3760:		vpmulld	%xmm15, %xmm1, %xmm1
>  3765:		vpaddd	%xmm15, %xmm1, %xmm3
>  3770:		vpandn	%xmm3, %xmm7, %xmm3
>  3774:		vpaddd	%xmm0, %xmm1, %xmm7
>  3778:		vpextrq	$1, %xmm7, %rsi
>  3784:		vmovq	%xmm7, %rbx
>  3789:		vpaddd	%xmm0, %xmm3, %xmm0
>  3793:		vpextrq	$1, %xmm0, %r8
>  3799:		vmovq	%xmm0, %r9
>  3804:		vpaddd	%xmm5, %xmm1, %xmm0
>  3808:		vpextrq	$1, %xmm0, %rdx
>  3814:		vmovq	%xmm0, %rdi
>  3819:		vpaddd	%xmm5, %xmm3, %xmm0
>  3823:		vpextrq	$1, %xmm0, %r10
>  3829:		vmovq	%xmm0, %r11
>  3834:		vpand	%xmm9, %xmm2, %xmm0
>  3839:		vpand	%xmm9, %xmm6, %xmm1
>  3844:		vpshufb	%xmm12, %xmm0, %xmm6
>  3849:		vpshufb	%xmm12, %xmm1, %xmm4
>  3854:		movslq	%ebx, %r14
>  3857:		sarq	$32, %rbx
>  3861:		movslq	%esi, %rcx
>  3864:		sarq	$32, %rsi
>  3868:		vmovd	(%r14,%rax), %xmm1
>  3874:		vpinsrd	$1, (%rbx,%rax), %xmm1, %xmm1
>  3881:		vpinsrd	$2, (%rcx,%rax), %xmm1, %xmm1
>  3888:		vpinsrd	$3, (%rsi,%rax), %xmm1, %xmm2
>  3895:		movslq	%edi, %rcx
>  3898:		sarq	$32, %rdi
>  3902:		movslq	%edx, %rsi
>  3905:		sarq	$32, %rdx
>  3909:		vmovd	(%rcx,%rax), %xmm1
>  3914:		vpinsrd	$1, (%rdi,%rax), %xmm1, %xmm1
>  3921:		vpinsrd	$2, (%rsi,%rax), %xmm1, %xmm1
>  3928:		vpinsrd	$3, (%rdx,%rax), %xmm1, %xmm7
>  3935:		vpmovzxbw	%xmm2, %xmm1
>  3940:		vpmovzxbw	%xmm7, %xmm3
>  3945:		vpsubw	%xmm1, %xmm3, %xmm3
>  3949:		vpmovzxbw	%xmm6, %xmm5
>  3954:		vpmullw	%xmm5, %xmm3, %xmm3
>  3958:		vpsrlw	$8, %xmm3, %xmm3
>  3963:		vpaddw	%xmm1, %xmm3, %xmm1
>  3967:		vpunpckhbw	%xmm10, %xmm6, %xmm3
>  3972:		vpunpckhbw	%xmm10, %xmm2, %xmm2
>  3977:		vpunpckhbw	%xmm10, %xmm7, %xmm6
>  3982:		vpsubw	%xmm2, %xmm6, %xmm6
>  3986:		vpmullw	%xmm3, %xmm6, %xmm6
>  3990:		vpsrlw	$8, %xmm6, %xmm6
>  3995:		vpaddw	%xmm2, %xmm6, %xmm2
>  3999:		vpand	%xmm8, %xmm1, %xmm1
>  4004:		vpand	%xmm8, %xmm2, %xmm2
>  4009:		vpackuswb	%xmm2, %xmm1, %xmm1
>  4013:		movslq	%r9d, %rcx
>  4016:		sarq	$32, %r9
>  4020:		movslq	%r8d, %rdx
>  4023:		sarq	$32, %r8
>  4027:		vmovd	(%rcx,%rax), %xmm2
>  4032:		vpinsrd	$1, (%r9,%rax), %xmm2, %xmm2
>  4039:		vpinsrd	$2, (%rdx,%rax), %xmm2, %xmm2
>  4046:		vpinsrd	$3, (%r8,%rax), %xmm2, %xmm2
>  4053:		movslq	%r11d, %rcx
>  4056:		sarq	$32, %r11
>  4060:		movslq	%r10d, %rdx
>  4063:		sarq	$32, %r10
>  4067:		vmovd	(%rcx,%rax), %xmm6
>  4072:		vpinsrd	$1, (%r11,%rax), %xmm6, %xmm6
>  4079:		vpinsrd	$2, (%rdx,%rax), %xmm6, %xmm6
>  4086:		vpinsrd	$3, (%r10,%rax), %xmm6, %xmm6
>  4093:		vpmovzxbw	%xmm2, %xmm7
>  4098:		vpmovzxbw	%xmm6, %xmm0
>  4103:		vpsubw	%xmm7, %xmm0, %xmm0
>  4107:		vpmullw	%xmm5, %xmm0, %xmm0
>  4111:		vpsrlw	$8, %xmm0, %xmm0
>  4116:		vpaddw	%xmm7, %xmm0, %xmm0
>  4120:		vpunpckhbw	%xmm10, %xmm2, %xmm2
>  4125:		vpunpckhbw	%xmm10, %xmm6, %xmm5
>  4130:		vpsubw	%xmm2, %xmm5, %xmm5
>  4134:		vpmullw	%xmm3, %xmm5, %xmm3
>  4138:		vpsrlw	$8, %xmm3, %xmm3
>  4143:		vpaddw	%xmm2, %xmm3, %xmm2
>  4147:		vpand	%xmm8, %xmm0, %xmm0
>  4152:		vpand	%xmm8, %xmm2, %xmm2
>  4157:		vpackuswb	%xmm2, %xmm0, %xmm0
>  4161:		vpmovzxbw	%xmm4, %xmm2
>  4166:		vpunpckhbw	%xmm10, %xmm4, %xmm3
>  4171:		vpmovzxbw	%xmm1, %xmm5
>  4176:		vpunpckhbw	%xmm10, %xmm1, %xmm1
>  4181:		vpunpckhbw	%xmm10, %xmm0, %xmm6
>  4186:		vpmovzxbw	%xmm0, %xmm0
>  4191:		vpsubw	%xmm5, %xmm0, %xmm0
>  4195:		vpmullw	%xmm2, %xmm0, %xmm0
>  4199:		vpsrlw	$8, %xmm0, %xmm0
>  4204:		vpaddw	%xmm5, %xmm0, %xmm0
>  4208:		vpsubw	%xmm1, %xmm6, %xmm2
>  4212:		vpmullw	%xmm3, %xmm2, %xmm2
>  4216:		vpsrlw	$8, %xmm2, %xmm2
>  4221:		vpaddw	%xmm1, %xmm2, %xmm1
>  4225:		vpand	%xmm8, %xmm0, %xmm0
>  4230:		vpand	%xmm8, %xmm1, %xmm1
>  4235:		vpackuswb	%xmm1, %xmm0, %xmm0
>  4239:		vpand	%xmm9, %xmm13, %xmm1
>  4244:		vcvtdq2ps	%xmm1, %xmm1
>  4248:		movabsq	$140164720566364, %rax
>  4258:		vbroadcastss	(%rax), %xmm2
>  4263:		vmulps	%xmm2, %xmm1, %xmm8
>  4267:		vpsrld	$8, %xmm13, %xmm3
>  4273:		vpand	%xmm9, %xmm3, %xmm3
>  4278:		vcvtdq2ps	%xmm3, %xmm3
>  4282:		vmulps	%xmm2, %xmm3, %xmm10
>  4286:		vpsrld	$16, %xmm13, %xmm5
>  4292:		vpand	%xmm9, %xmm5, %xmm5
>  4297:		vcvtdq2ps	%xmm5, %xmm5
>  4301:		vmulps	%xmm2, %xmm5, %xmm5
>  4305:		vpsrld	$24, %xmm13, %xmm6
>  4311:		vcvtdq2ps	%xmm6, %xmm6
>  4315:		vmulps	%xmm2, %xmm6, %xmm6
>  4319:		vpsrld	$8, %xmm0, %xmm7
>  4324:		vpand	%xmm9, %xmm7, %xmm7
>  4329:		vpsrld	$16, %xmm0, %xmm1
>  4334:		vpand	%xmm9, %xmm1, %xmm1
>  4339:		vpand	%xmm9, %xmm0, %xmm3
>  4344:		vcvtdq2ps	%xmm3, %xmm3
>  4348:		vmulps	%xmm2, %xmm3, %xmm3
>  4352:		vcvtdq2ps	%xmm7, %xmm7
>  4356:		vmulps	%xmm2, %xmm7, %xmm7
>  4360:		vcvtdq2ps	%xmm1, %xmm1
>  4364:		vmulps	%xmm2, %xmm1, %xmm1
>  4368:		vpsrld	$24, %xmm0, %xmm0
>  4373:		vcvtdq2ps	%xmm0, %xmm0
>  4377:		vmulps	%xmm2, %xmm0, %xmm0
>  4381:		vinsertf128	$1, %xmm3, %ymm8, %ymm2
>  4387:		vinsertf128	$1, %xmm7, %ymm10, %ymm3
>  4393:		vinsertf128	$1, %xmm1, %ymm5, %ymm1
>  4399:		vinsertf128	$1, %xmm0, %ymm6, %ymm0
>  4405:		vmovaps	640(%rsp), %ymm5
>  4414:		vblendvps	%ymm11, %ymm2, %ymm5, %ymm2
>  4420:		vblendvps	%ymm11, %ymm3, %ymm14, %ymm3
>  4426:		vmovaps	416(%rsp), %ymm5
>  4435:		vblendvps	%ymm11, %ymm1, %ymm5, %ymm1
>  4441:		vmovaps	320(%rsp), %ymm5
>  4450:		vmulps	%ymm2, %ymm5, %ymm2
>  4454:		vblendvps	%ymm11, %ymm2, %ymm5, %ymm2
>  4460:		vmovaps	352(%rsp), %ymm5
>  4469:		vmulps	%ymm3, %ymm5, %ymm3
>  4473:		vblendvps	%ymm11, %ymm3, %ymm5, %ymm3
>  4479:		vmovaps	288(%rsp), %ymm5
>  4488:		vmulps	%ymm1, %ymm5, %ymm1
>  4492:		vblendvps	%ymm11, %ymm1, %ymm5, %ymm1
>  4498:		vxorps	%ymm8, %ymm8, %ymm8
>  4503:		vblendvps	%ymm11, %ymm0, %ymm8, %ymm0
>  4509:		vmovaps	256(%rsp), %ymm5
>  4518:		vmulps	%ymm0, %ymm5, %ymm0
>  4522:		vblendvps	%ymm11, %ymm0, %ymm5, %ymm0
>  4528:		vcmpleps	%ymm8, %ymm0, %ymm5
>  4534:		vmovaps	448(%rsp), %ymm6
>  4543:		vblendvps	%ymm11, %ymm5, %ymm6, %ymm5
>  4549:		vandps	384(%rsp), %ymm5, %ymm6
>  4558:		vblendvps	%ymm11, %ymm6, %ymm5, %ymm5
>  4564:		vmovaps	608(%rsp), %ymm6
>  4573:		vxorps	%ymm6, %ymm5, %ymm5
>  4577:		vcmpnltps	%ymm8, %ymm5, %ymm5
>  4583:		vpcmpeqd	%ymm6, %ymm6, %ymm6
>  4587:		vpxor	%ymm6, %ymm11, %ymm6
>  4591:		vorps	%ymm6, %ymm5, %ymm5
>  4595:		movq	%r15, %rax
>  4598:		vmovaps	288(%rax), %ymm6
>  4606:		vandps	%ymm6, %ymm5, %ymm5
>  4610:		vmovaps	512(%rsp), %ymm6
>  4619:		vblendvps	%ymm11, 480(%rsp), %ymm6, %ymm6
>  4630:		vmovaps	576(%rsp), %ymm7
>  4639:		vblendvps	%ymm11, 544(%rsp), %ymm7, %ymm7
>  4650:		vblendvps	%ymm11, %ymm2, %ymm8, %ymm2
>  4656:		vblendvps	%ymm11, %ymm3, %ymm8, %ymm3
>  4662:		vblendvps	%ymm11, %ymm1, %ymm8, %ymm1
>  4668:		vblendvps	%ymm11, %ymm0, %ymm8, %ymm0
>  4674:		vmovaps	%ymm2, 832(%rax)
>  4682:		vmovaps	%ymm3, 864(%rax)
>  4690:		vmovaps	%ymm1, 896(%rax)
>  4698:		vmovaps	%ymm0, 928(%rax)
>  4706:		vmovaps	%ymm6, 960(%rax)
>  4714:		vmovaps	%ymm7, 992(%rax)
>  4722:		vmovaps	%ymm8, 1024(%rax)
>  4730:		vmovaps	%ymm8, 1056(%rax)
>  4738:		vmovaps	%ymm5, 288(%rax)
>  4746:		leaq	-40(%rbp), %rsp
>  4750:		popq	%rbx
>  4751:		popq	%r12
>  4753:		popq	%r13
>  4755:		popq	%r14
>  4757:		popq	%r15
>  4759:		popq	%rbp
>  4760:		vzeroupper
>  4763:		retq
>
>
>frag shader  0x7f7aa4613000
>BLEND shader 0x7f7aa4611000
>BLEND shader 0x7f7aa460f000
>fetch shader 0x7f7aa460d000
>
>=========================================================
>Process id 28383 Caught SIGSEGV at 0x5ad13fc address not mapped to object
>Program Stack:
>WARNING: The stack trace will not use advanced capabilities because this is a release build.
>0x31aae326a0 : ??? [(???) ???:-1]
>0x7f7aa4613c71 : ??? [(???) ???:-1]
>0x7f7aa47b9bdb : ??? [(???) ???:-1]
>0x7f7aa4906a88 : ??? [(???) ???:-1]
>0x7f7aa4914b74 : ??? [(???) ???:-1]
>0x7f7aa47910f5 : ??? [(???) ???:-1]
>0x7f7aa4790604 : ??? [(???) ???:-1]
>0x7f7aa478f954 : ??? [(???) ???:-1]
>0x7f7aa478fa34 : ??? [(???) ???:-1]
>0x7f7aa477422a : ??? [(???) ???:-1]
>0x7f7aa935a705 : ??? [(???) ???:-1]
>0x7f7aa91c7768 : ??? [(???) ???:-1]
>0x7f7aa916e0b5 : ??? [(???) ???:-1]
>0x7f7aa916e2f1 : ??? [(???) ???:-1]
>0x7f7aa916e3ca : ??? [(???) ???:-1]
>0x7f7aa8c4dd0b : shared_dispatch_stub_338 [(libglapi.so.0) ???:-1]
>0x7f7aa9eaacb8 : vtkOpenGLPolyDataMapper::RenderPieceDraw(vtkRenderer*, vtkActor*) [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9eaa6e1 : vtkOpenGLPolyDataMapper::RenderPiece(vtkRenderer*, vtkActor*) [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aadf1717a : vtkPolyDataMapper::Render(vtkRenderer*, vtkActor*) [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aa9e4c0f8 : vtkOpenGLActor::Render(vtkRenderer*, vtkMapper*) [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aade9afaa : vtkActor::RenderTranslucentPolygonalGeometry(vtkViewport*) [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aadf1bdf5 : vtkProp::RenderFilteredTranslucentPolygonalGeometry(vtkViewport*, vtkInformation*) [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aa9e390cf : vtkDefaultPass::RenderFilteredTranslucentPolygonalGeometry(vtkRenderState const*) [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9e3d768 : vtkDualDepthPeelingPass::RenderTranslucentPass() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9e3deae : vtkDualDepthPeelingPass::InitializeDepth() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9e3e07e : vtkDualDepthPeelingPass::Prepare() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9e3ea94 : vtkDualDepthPeelingPass::Render(vtkRenderState const*) [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9ec9651 : vtkOpenGLRenderer::DeviceRenderTranslucentPolygonalGeometry() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9ec9a30 : vtkOpenGLRenderer::UpdateGeometry() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9ec9b28 : vtkOpenGLRenderer::DeviceRender() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aadf254ae : vtkRenderer::Render() [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aadf2181c : vtkRendererCollection::Render() [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aadf29ff5 : vtkRenderWindow::DoStereoRender() [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aadf2b36d : vtkRenderWindow::DoFDRender() [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aadf2bcc5 : vtkRenderWindow::DoAARender() [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aadf2a36b : vtkRenderWindow::Render() [(libvtkRenderingCore-7.1.so.1) ???:-1]
>0x7f7aa9ec87a1 : vtkOpenGLRenderWindow::Render() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x7f7aa9f38f90 : vtkXOpenGLRenderWindow::Render() [(libvtkRenderingOpenGL2-7.1.so.1) ???:-1]
>0x43340d : TestTextureRGBADepthPeeling(int, char**) [(vtkRenderingCoreCxxTests) ???:-1]
>0x40fb71 : main [(vtkRenderingCoreCxxTests) ???:-1]
>0x31aae1ed5d : __libc_start_main [(libc.so.6) ???:-1]
>0x411df1 : ??? [(???) ???:-1]
>=========================================================
>
>1/1 Test #vtkRenderingCoreCxx-TestTextureRGBADepthPeeling ...***Exception: Other  1.66 sec
>
>0% tests passed, 1 tests failed out of 1
>
>Label Time Summary:
>vtkRenderingCore    =   1.66 sec
>
>Total Test time (real) =   1.85 sec
>
>The following tests FAILED:
>	1525 - vtkRenderingCoreCxx-TestTextureRGBADepthPeeling (OTHER_FAULT)
>Errors while running CTest

Actions: View

Attachments on bug 94955: 122975 | 123134 | 123136 | 123137

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.