Attachment 123134 Details for Bug 94955 – tex_sample_func

tex_sample_func

tex_sample_func-llvmpipe.txt (text/plain), 11.37 KB, created by Bruce Cherniak on 2016-04-21 19:36:31 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Bruce Cherniak

Created: 2016-04-21 19:36:31 UTC

Size: 11.37 KB

patch

obsolete

>define private fastcc { <8 x float>, <8 x float>, <8 x float>, <8 x float> } @texfunc_res_0_sam_0_0({ [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* noalias, <8 x float>, <8 x float>) {
>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>
>  %3 = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 0
>  %context.texture0.width = load i32, i32* %3
>  %context.texture0.row_stride = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 6
>  %context.texture0.img_stride = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 7
>  %4 = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 5
>  %context.texture0.base_ptr = load i8*, i8** %4
>  %context.texture0.mip_offsets = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 8
>  %5 = insertelement <4 x i32> undef, i32 %context.texture0.width, i32 0
>  %6 = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 1
>  %context.texture0.height = load i32, i32* %6
>  %7 = insertelement <4 x i32> %5, i32 %context.texture0.height, i32 1
>  %8 = getelementptr { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }, { [16 x float*], [16 x i32], float, i32, i32, i8*, float*, { float, float }*, [32 x { i32, i32, i32, i32, i32, i8*, [14 x i32], [14 x i32], [14 x i32] }], [32 x { float, float, float, [4 x float] }] }* %0, i32 0, i32 8, i32 0, i32 3
>  %context.texture0.first_level = load i32, i32* %8
>  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
>  %9 = insertelement <4 x i32> undef, i32 %context.texture0.first_level, i32 0
>  %10 = shufflevector <4 x i32> %9, <4 x i32> undef, <4 x i32> zeroinitializer
>  %minify = lshr <4 x i32> %7, %10
>  %11 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %minify, <4 x i32> <i32 1, i32 1, i32 1, i32 1>)
>  %12 = getelementptr [14 x i32], [14 x i32]* %context.texture0.row_stride, i32 0, i32 %context.texture0.first_level
>  %13 = load i32, i32* %12
>  %14 = insertelement <8 x i32> undef, i32 %13, i32 0
>  %15 = shufflevector <8 x i32> %14, <8 x i32> undef, <8 x i32> zeroinitializer
>  %16 = getelementptr [14 x i32], [14 x i32]* %context.texture0.mip_offsets, i32 0, i32 %context.texture0.first_level
>  %17 = load i32, i32* %16
>  %18 = getelementptr i8, i8* %context.texture0.base_ptr, i32 %17
>  %19 = shufflevector <4 x i32> %11, <4 x i32> undef, <8 x i32> zeroinitializer
>  %20 = shufflevector <4 x i32> %11, <4 x i32> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %21 = sitofp <4 x i32> %11 to <4 x float>
>  %22 = shufflevector <4 x float> %21, <4 x float> undef, <8 x i32> zeroinitializer
>  %23 = shufflevector <4 x float> %21, <4 x float> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %24 = sub <8 x i32> %19, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %25 = fmul <8 x float> %1, %22
>  %26 = fptosi <8 x float> %25 to <8 x i32>
>  %27 = shufflevector <8 x i32> %26, <8 x i32> %26, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %28 = shufflevector <8 x i32> %24, <8 x i32> %24, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %29 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %27, <4 x i32> %28)
>  %30 = shufflevector <8 x i32> %26, <8 x i32> %26, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %31 = shufflevector <8 x i32> %24, <8 x i32> %24, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %32 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %30, <4 x i32> %31)
>  %33 = shufflevector <4 x i32> %29, <4 x i32> %32, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %34 = shufflevector <8 x i32> %33, <8 x i32> %33, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %35 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %34, <4 x i32> zeroinitializer)
>  %36 = shufflevector <8 x i32> %33, <8 x i32> %33, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %37 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %36, <4 x i32> zeroinitializer)
>  %tex.x.wrapped = shufflevector <4 x i32> %35, <4 x i32> %37, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %38 = sub <8 x i32> %20, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
>  %39 = fmul <8 x float> %2, %23
>  %40 = fptosi <8 x float> %39 to <8 x i32>
>  %41 = shufflevector <8 x i32> %40, <8 x i32> %40, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %42 = shufflevector <8 x i32> %38, <8 x i32> %38, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %43 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %41, <4 x i32> %42)
>  %44 = shufflevector <8 x i32> %40, <8 x i32> %40, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %45 = shufflevector <8 x i32> %38, <8 x i32> %38, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %46 = call <4 x i32> @llvm.x86.sse41.pminsd(<4 x i32> %44, <4 x i32> %45)
>  %47 = shufflevector <4 x i32> %43, <4 x i32> %46, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %48 = shufflevector <8 x i32> %47, <8 x i32> %47, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
>  %49 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %48, <4 x i32> zeroinitializer)
>  %50 = shufflevector <8 x i32> %47, <8 x i32> %47, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
>  %51 = call <4 x i32> @llvm.x86.sse41.pmaxsd(<4 x i32> %50, <4 x i32> zeroinitializer)
>  %tex.y.wrapped = shufflevector <4 x i32> %49, <4 x i32> %51, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
>  %52 = mul <8 x i32> %tex.x.wrapped, <i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4>
>  %53 = mul <8 x i32> %tex.y.wrapped, %15
>  %54 = add <8 x i32> %52, %53
>  %55 = extractelement <8 x i32> %54, i32 0
>  %56 = getelementptr i8, i8* %18, i32 %55
>  %57 = bitcast i8* %56 to i32*
>  %58 = load i32, i32* %57
>  %59 = insertelement <8 x i32> undef, i32 %58, i32 0
>  %60 = extractelement <8 x i32> %54, i32 1
>  %61 = getelementptr i8, i8* %18, i32 %60
>  %62 = bitcast i8* %61 to i32*
>  %63 = load i32, i32* %62
>  %64 = insertelement <8 x i32> %59, i32 %63, i32 1
>  %65 = extractelement <8 x i32> %54, i32 2
>  %66 = getelementptr i8, i8* %18, i32 %65
>  %67 = bitcast i8* %66 to i32*
>  %68 = load i32, i32* %67
>  %69 = insertelement <8 x i32> %64, i32 %68, i32 2
>  %70 = extractelement <8 x i32> %54, i32 3
>  %71 = getelementptr i8, i8* %18, i32 %70
>  %72 = bitcast i8* %71 to i32*
>  %73 = load i32, i32* %72
>  %74 = insertelement <8 x i32> %69, i32 %73, i32 3
>  %75 = extractelement <8 x i32> %54, i32 4
>  %76 = getelementptr i8, i8* %18, i32 %75
>  %77 = bitcast i8* %76 to i32*
>  %78 = load i32, i32* %77
>  %79 = insertelement <8 x i32> %74, i32 %78, i32 4
>  %80 = extractelement <8 x i32> %54, i32 5
>  %81 = getelementptr i8, i8* %18, i32 %80
>  %82 = bitcast i8* %81 to i32*
>  %83 = load i32, i32* %82
>  %84 = insertelement <8 x i32> %79, i32 %83, i32 5
>  %85 = extractelement <8 x i32> %54, i32 6
>  %86 = getelementptr i8, i8* %18, i32 %85
>  %87 = bitcast i8* %86 to i32*
>  %88 = load i32, i32* %87
>  %89 = insertelement <8 x i32> %84, i32 %88, i32 6
>  %90 = extractelement <8 x i32> %54, i32 7
>  %91 = getelementptr i8, i8* %18, i32 %90
>  %92 = bitcast i8* %91 to i32*
>  %93 = load i32, i32* %92
>  %94 = insertelement <8 x i32> %89, i32 %93, i32 7
>  %95 = and <8 x i32> %94, <i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215, i32 16777215>
>  %96 = sitofp <8 x i32> %95 to <8 x float>
>  %97 = fmul <8 x float> %96, <float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000, float 0x3E70000020000000>
>  store <8 x float> %97, <8 x float>* %sampler0_texel_x_var
>  store <8 x float> %97, <8 x float>* %sampler0_texel_y_var
>  store <8 x float> %97, <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
>  %mrv = insertvalue { <8 x float>, <8 x float>, <8 x float>, <8 x float> } undef, <8 x float> %sampler0_texel_x, 0
>  %mrv1 = insertvalue { <8 x float>, <8 x float>, <8 x float>, <8 x float> } %mrv, <8 x float> %sampler0_texel_y, 1
>  %mrv2 = insertvalue { <8 x float>, <8 x float>, <8 x float>, <8 x float> } %mrv1, <8 x float> %sampler0_texel_z, 2
>  %mrv3 = insertvalue { <8 x float>, <8 x float>, <8 x float>, <8 x float> } %mrv2, <8 x float> %sampler0_texel_w, 3
>  ret { <8 x float>, <8 x float>, <8 x float>, <8 x float> } %mrv3
>}

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.