debug_get_bool_option: SWRAST_NO_PRESENT = FALSE
debug_get_option: GALLIUM_DRIVER = llvmpipe
debug_get_flags_option: LP_DEBUG = 0x0
debug_get_flags_option: GALLIVM_DEBUG = 0x0
debug_get_bool_option: GALLIUM_DUMP_CPU = TRUE
util_cpu_caps.arch = 4
util_cpu_caps.nr_cpus = 1
util_cpu_caps.x86_cpu_type = 8
util_cpu_caps.cacheline = 64
util_cpu_caps.has_tsc = 1
util_cpu_caps.has_mmx = 1
util_cpu_caps.has_mmx2 = 1
util_cpu_caps.has_sse = 1
util_cpu_caps.has_sse2 = 1
util_cpu_caps.has_sse3 = 1
util_cpu_caps.has_ssse3 = 0
util_cpu_caps.has_sse4_1 = 0
util_cpu_caps.has_3dnow = 1
util_cpu_caps.has_3dnow_ext = 1
util_cpu_caps.has_altivec = 0
debug_get_num_option: LP_NUM_THREADS = 1
couldn't open libtxc_dxtn.so, software DXTn compression/decompression unavailable
debug_get_bool_option: GALLIUM_RBUG = FALSE
debug_get_option: GALLIUM_TRACE = (null)
debug_get_bool_option: DRAW_FSE = FALSE
debug_get_bool_option: DRAW_NO_FSE = FALSE
debug_get_bool_option: DRAW_USE_LLVM = TRUE
debug_get_bool_option: GALLIUM_DUMP_VS = FALSE
debug_get_bool_option: LP_NO_RAST = FALSE
debug_get_bool_option: MESA_MVP_DP4 = FALSE
debug_get_flags_option: ST_DEBUG = 0x0
debug_get_bool_option: TGSI_PRINT_SANITY = TRUE
Floating-point arithmetic operators only work with floating-point types!
  %24 = fadd i32 %start, %count                   ; <i32> [#uses=1]
Broken module found, verification continues.

define void @draw_llvm_shader(%draw_jit_context* noalias %context, %vertex_header3* noalias %io, i8** noalias %vbuffers, i32 %start, i32 %count, i32 %stride, %pipe_vertex_buffer* noalias %vb, i32 %instance_id) {
entry:
  %0 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %1 = alloca <4 x float>                         ; <<4 x float>*> [#uses=3]
  %2 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %3 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %4 = alloca <4 x float>                         ; <<4 x float>*> [#uses=6]
  %5 = alloca <4 x float>                         ; <<4 x float>*> [#uses=7]
  %6 = alloca <4 x float>                         ; <<4 x float>*> [#uses=7]
  %7 = alloca <4 x float>                         ; <<4 x float>*> [#uses=7]
  %8 = alloca <4 x float>                         ; <<4 x float>*> [#uses=1]
  %9 = alloca <4 x float>                         ; <<4 x float>*> [#uses=5]
  %10 = alloca <4 x float>                        ; <<4 x float>*> [#uses=5]
  %11 = alloca <4 x float>                        ; <<4 x float>*> [#uses=8]
  %12 = alloca <4 x float>                        ; <<4 x float>*> [#uses=7]
  %13 = alloca <4 x float>                        ; <<4 x float>*> [#uses=8]
  %14 = alloca <4 x float>                        ; <<4 x float>*> [#uses=8]
  %15 = alloca <4 x float>                        ; <<4 x float>*> [#uses=8]
  %16 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %17 = alloca <4 x float>                        ; <<4 x float>*> [#uses=3]
  %18 = alloca <4 x float>                        ; <<4 x float>*> [#uses=3]
  %19 = alloca <4 x float>                        ; <<4 x float>*> [#uses=3]
  %20 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %21 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %22 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %23 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %24 = fadd i32 %start, %count                   ; <i32> [#uses=1]
  br label %loop

loop:                                             ; preds = %loop, %entry
  %25 = phi i32 [ %start, %entry ], [ %578, %loop ] ; <i32> [#uses=6]
  %26 = sub i32 %25, %start                       ; <i32> [#uses=1]
  %27 = getelementptr %vertex_header3* %io, i32 %26 ; <%vertex_header3*> [#uses=8]
  %28 = add i32 %25, 0                            ; <i32> [#uses=2]
  %29 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %30 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr = getelementptr %pipe_vertex_buffer* %29, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride = load i32* %.stride_ptr               ; <i32> [#uses=1]
  %.max_index_ptr = getelementptr %pipe_vertex_buffer* %29, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index = load i32* %.max_index_ptr         ; <i32> [#uses=2]
  %.buffer_offset_ptr = getelementptr %pipe_vertex_buffer* %29, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset = load i32* %.buffer_offset_ptr ; <i32> [#uses=1]
  %31 = icmp ule i32 %28, %.max_index             ; <i1> [#uses=1]
  %32 = select i1 %31, i32 %28, i32 %.max_index   ; <i32> [#uses=1]
  %33 = mul i32 %.stride, %32                     ; <i32> [#uses=1]
  %vbuffer = load i8** %30                        ; <i8*> [#uses=1]
  %34 = add i32 %33, %.buffer_offset              ; <i32> [#uses=1]
  %35 = add i32 %34, 0                            ; <i32> [#uses=1]
  %36 = getelementptr i8* %vbuffer, i32 %35       ; <i8*> [#uses=3]
  %src_tmp = getelementptr i8* %36, i32 0         ; <i8*> [#uses=1]
  %37 = bitcast i8* %src_tmp to float*            ; <float*> [#uses=1]
  %38 = load float* %37                           ; <float> [#uses=1]
  %39 = insertelement <4 x float> zeroinitializer, float %38, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp1 = getelementptr i8* %36, i32 4        ; <i8*> [#uses=1]
  %40 = bitcast i8* %src_tmp1 to float*           ; <float*> [#uses=1]
  %41 = load float* %40                           ; <float> [#uses=1]
  %42 = insertelement <4 x float> %39, float %41, i32 1 ; <<4 x float>> [#uses=1]
  %src_tmp2 = getelementptr i8* %36, i32 8        ; <i8*> [#uses=1]
  %43 = bitcast i8* %src_tmp2 to float*           ; <float*> [#uses=1]
  %44 = load float* %43                           ; <float> [#uses=1]
  %45 = insertelement <4 x float> %42, float %44, i32 2 ; <<4 x float>> [#uses=1]
  %46 = insertelement <4 x float> %45, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %47 = add i32 %25, 1                            ; <i32> [#uses=2]
  %48 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %49 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr3 = getelementptr %pipe_vertex_buffer* %48, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride4 = load i32* %.stride_ptr3             ; <i32> [#uses=1]
  %.max_index_ptr5 = getelementptr %pipe_vertex_buffer* %48, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index6 = load i32* %.max_index_ptr5       ; <i32> [#uses=2]
  %.buffer_offset_ptr7 = getelementptr %pipe_vertex_buffer* %48, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset8 = load i32* %.buffer_offset_ptr7 ; <i32> [#uses=1]
  %50 = icmp ule i32 %47, %.max_index6            ; <i1> [#uses=1]
  %51 = select i1 %50, i32 %47, i32 %.max_index6  ; <i32> [#uses=1]
  %52 = mul i32 %.stride4, %51                    ; <i32> [#uses=1]
  %vbuffer9 = load i8** %49                       ; <i8*> [#uses=1]
  %53 = add i32 %52, %.buffer_offset8             ; <i32> [#uses=1]
  %54 = add i32 %53, 0                            ; <i32> [#uses=1]
  %55 = getelementptr i8* %vbuffer9, i32 %54      ; <i8*> [#uses=3]
  %src_tmp10 = getelementptr i8* %55, i32 0       ; <i8*> [#uses=1]
  %56 = bitcast i8* %src_tmp10 to float*          ; <float*> [#uses=1]
  %57 = load float* %56                           ; <float> [#uses=1]
  %58 = insertelement <4 x float> zeroinitializer, float %57, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp11 = getelementptr i8* %55, i32 4       ; <i8*> [#uses=1]
  %59 = bitcast i8* %src_tmp11 to float*          ; <float*> [#uses=1]
  %60 = load float* %59                           ; <float> [#uses=1]
  %61 = insertelement <4 x float> %58, float %60, i32 1 ; <<4 x float>> [#uses=1]
  %src_tmp12 = getelementptr i8* %55, i32 8       ; <i8*> [#uses=1]
  %62 = bitcast i8* %src_tmp12 to float*          ; <float*> [#uses=1]
  %63 = load float* %62                           ; <float> [#uses=1]
  %64 = insertelement <4 x float> %61, float %63, i32 2 ; <<4 x float>> [#uses=1]
  %65 = insertelement <4 x float> %64, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %66 = add i32 %25, 2                            ; <i32> [#uses=2]
  %67 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %68 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr13 = getelementptr %pipe_vertex_buffer* %67, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride14 = load i32* %.stride_ptr13           ; <i32> [#uses=1]
  %.max_index_ptr15 = getelementptr %pipe_vertex_buffer* %67, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index16 = load i32* %.max_index_ptr15     ; <i32> [#uses=2]
  %.buffer_offset_ptr17 = getelementptr %pipe_vertex_buffer* %67, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset18 = load i32* %.buffer_offset_ptr17 ; <i32> [#uses=1]
  %69 = icmp ule i32 %66, %.max_index16           ; <i1> [#uses=1]
  %70 = select i1 %69, i32 %66, i32 %.max_index16 ; <i32> [#uses=1]
  %71 = mul i32 %.stride14, %70                   ; <i32> [#uses=1]
  %vbuffer19 = load i8** %68                      ; <i8*> [#uses=1]
  %72 = add i32 %71, %.buffer_offset18            ; <i32> [#uses=1]
  %73 = add i32 %72, 0                            ; <i32> [#uses=1]
  %74 = getelementptr i8* %vbuffer19, i32 %73     ; <i8*> [#uses=3]
  %src_tmp20 = getelementptr i8* %74, i32 0       ; <i8*> [#uses=1]
  %75 = bitcast i8* %src_tmp20 to float*          ; <float*> [#uses=1]
  %76 = load float* %75                           ; <float> [#uses=1]
  %77 = insertelement <4 x float> zeroinitializer, float %76, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp21 = getelementptr i8* %74, i32 4       ; <i8*> [#uses=1]
  %78 = bitcast i8* %src_tmp21 to float*          ; <float*> [#uses=1]
  %79 = load float* %78                           ; <float> [#uses=1]
  %80 = insertelement <4 x float> %77, float %79, i32 1 ; <<4 x float>> [#uses=1]
  %src_tmp22 = getelementptr i8* %74, i32 8       ; <i8*> [#uses=1]
  %81 = bitcast i8* %src_tmp22 to float*          ; <float*> [#uses=1]
  %82 = load float* %81                           ; <float> [#uses=1]
  %83 = insertelement <4 x float> %80, float %82, i32 2 ; <<4 x float>> [#uses=1]
  %84 = insertelement <4 x float> %83, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %85 = add i32 %25, 3                            ; <i32> [#uses=2]
  %86 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %87 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr23 = getelementptr %pipe_vertex_buffer* %86, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride24 = load i32* %.stride_ptr23           ; <i32> [#uses=1]
  %.max_index_ptr25 = getelementptr %pipe_vertex_buffer* %86, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index26 = load i32* %.max_index_ptr25     ; <i32> [#uses=2]
  %.buffer_offset_ptr27 = getelementptr %pipe_vertex_buffer* %86, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset28 = load i32* %.buffer_offset_ptr27 ; <i32> [#uses=1]
  %88 = icmp ule i32 %85, %.max_index26           ; <i1> [#uses=1]
  %89 = select i1 %88, i32 %85, i32 %.max_index26 ; <i32> [#uses=1]
  %90 = mul i32 %.stride24, %89                   ; <i32> [#uses=1]
  %vbuffer29 = load i8** %87                      ; <i8*> [#uses=1]
  %91 = add i32 %90, %.buffer_offset28            ; <i32> [#uses=1]
  %92 = add i32 %91, 0                            ; <i32> [#uses=1]
  %93 = getelementptr i8* %vbuffer29, i32 %92     ; <i8*> [#uses=3]
  %src_tmp30 = getelementptr i8* %93, i32 0       ; <i8*> [#uses=1]
  %94 = bitcast i8* %src_tmp30 to float*          ; <float*> [#uses=1]
  %95 = load float* %94                           ; <float> [#uses=1]
  %96 = insertelement <4 x float> zeroinitializer, float %95, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp31 = getelementptr i8* %93, i32 4       ; <i8*> [#uses=1]
  %97 = bitcast i8* %src_tmp31 to float*          ; <float*> [#uses=1]
  %98 = load float* %97                           ; <float> [#uses=1]
  %99 = insertelement <4 x float> %96, float %98, i32 1 ; <<4 x float>> [#uses=1]
  %src_tmp32 = getelementptr i8* %93, i32 8       ; <i8*> [#uses=1]
  %100 = bitcast i8* %src_tmp32 to float*         ; <float*> [#uses=1]
  %101 = load float* %100                         ; <float> [#uses=1]
  %102 = insertelement <4 x float> %99, float %101, i32 2 ; <<4 x float>> [#uses=1]
  %103 = insertelement <4 x float> %102, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %104 = extractelement <4 x float> %46, i32 0    ; <float> [#uses=1]
  %105 = insertelement <4 x float> zeroinitializer, float %104, i32 0 ; <<4 x float>> [#uses=1]
  %106 = extractelement <4 x float> %65, i32 0    ; <float> [#uses=1]
  %107 = insertelement <4 x float> %105, float %106, i32 1 ; <<4 x float>> [#uses=1]
  %108 = extractelement <4 x float> %84, i32 0    ; <float> [#uses=1]
  %109 = insertelement <4 x float> %107, float %108, i32 2 ; <<4 x float>> [#uses=1]
  %110 = extractelement <4 x float> %103, i32 0   ; <float> [#uses=1]
  %111 = insertelement <4 x float> %109, float %110, i32 3 ; <<4 x float>> [#uses=4]
  %112 = extractelement <4 x float> %46, i32 1    ; <float> [#uses=1]
  %113 = insertelement <4 x float> zeroinitializer, float %112, i32 0 ; <<4 x float>> [#uses=1]
  %114 = extractelement <4 x float> %65, i32 1    ; <float> [#uses=1]
  %115 = insertelement <4 x float> %113, float %114, i32 1 ; <<4 x float>> [#uses=1]
  %116 = extractelement <4 x float> %84, i32 1    ; <float> [#uses=1]
  %117 = insertelement <4 x float> %115, float %116, i32 2 ; <<4 x float>> [#uses=1]
  %118 = extractelement <4 x float> %103, i32 1   ; <float> [#uses=1]
  %119 = insertelement <4 x float> %117, float %118, i32 3 ; <<4 x float>> [#uses=4]
  %120 = extractelement <4 x float> %46, i32 2    ; <float> [#uses=1]
  %121 = insertelement <4 x float> zeroinitializer, float %120, i32 0 ; <<4 x float>> [#uses=1]
  %122 = extractelement <4 x float> %65, i32 2    ; <float> [#uses=1]
  %123 = insertelement <4 x float> %121, float %122, i32 1 ; <<4 x float>> [#uses=1]
  %124 = extractelement <4 x float> %84, i32 2    ; <float> [#uses=1]
  %125 = insertelement <4 x float> %123, float %124, i32 2 ; <<4 x float>> [#uses=1]
  %126 = extractelement <4 x float> %103, i32 2   ; <float> [#uses=1]
  %127 = insertelement <4 x float> %125, float %126, i32 3 ; <<4 x float>> [#uses=4]
  %128 = extractelement <4 x float> %46, i32 3    ; <float> [#uses=1]
  %129 = insertelement <4 x float> zeroinitializer, float %128, i32 0 ; <<4 x float>> [#uses=1]
  %130 = extractelement <4 x float> %65, i32 3    ; <float> [#uses=1]
  %131 = insertelement <4 x float> %129, float %130, i32 1 ; <<4 x float>> [#uses=1]
  %132 = extractelement <4 x float> %84, i32 3    ; <float> [#uses=1]
  %133 = insertelement <4 x float> %131, float %132, i32 2 ; <<4 x float>> [#uses=1]
  %134 = extractelement <4 x float> %103, i32 3   ; <float> [#uses=1]
  %135 = insertelement <4 x float> %133, float %134, i32 3 ; <<4 x float>> [#uses=4]
  %context.vs_constants_ptr = getelementptr %draw_jit_context* %context, i32 0, i32 0 ; <float**> [#uses=1]
  %context.vs_constants = load float** %context.vs_constants_ptr ; <float*> [#uses=44]
  %136 = getelementptr float* %context.vs_constants, i32 0 ; <float*> [#uses=1]
  %137 = load float* %136                         ; <float> [#uses=1]
  %138 = insertelement <4 x float> undef, float %137, i32 0 ; <<4 x float>> [#uses=1]
  %139 = shufflevector <4 x float> %138, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %140 = fmul <4 x float> %111, %139              ; <<4 x float>> [#uses=1]
  %141 = getelementptr float* %context.vs_constants, i32 1 ; <float*> [#uses=1]
  %142 = load float* %141                         ; <float> [#uses=1]
  %143 = insertelement <4 x float> undef, float %142, i32 0 ; <<4 x float>> [#uses=1]
  %144 = shufflevector <4 x float> %143, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %145 = fmul <4 x float> %111, %144              ; <<4 x float>> [#uses=1]
  %146 = getelementptr float* %context.vs_constants, i32 2 ; <float*> [#uses=1]
  %147 = load float* %146                         ; <float> [#uses=1]
  %148 = insertelement <4 x float> undef, float %147, i32 0 ; <<4 x float>> [#uses=1]
  %149 = shufflevector <4 x float> %148, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %150 = fmul <4 x float> %111, %149              ; <<4 x float>> [#uses=1]
  %151 = getelementptr float* %context.vs_constants, i32 3 ; <float*> [#uses=1]
  %152 = load float* %151                         ; <float> [#uses=1]
  %153 = insertelement <4 x float> undef, float %152, i32 0 ; <<4 x float>> [#uses=1]
  %154 = shufflevector <4 x float> %153, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %155 = fmul <4 x float> %111, %154              ; <<4 x float>> [#uses=1]
  store <4 x float> %140, <4 x float>* %15
  store <4 x float> %145, <4 x float>* %14
  store <4 x float> %150, <4 x float>* %13
  store <4 x float> %155, <4 x float>* %12
  %156 = getelementptr float* %context.vs_constants, i32 4 ; <float*> [#uses=1]
  %157 = load float* %156                         ; <float> [#uses=1]
  %158 = insertelement <4 x float> undef, float %157, i32 0 ; <<4 x float>> [#uses=1]
  %159 = shufflevector <4 x float> %158, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %160 = load <4 x float>* %15                    ; <<4 x float>> [#uses=1]
  %161 = fmul <4 x float> %119, %159              ; <<4 x float>> [#uses=1]
  %162 = fadd <4 x float> %161, %160              ; <<4 x float>> [#uses=1]
  %163 = getelementptr float* %context.vs_constants, i32 5 ; <float*> [#uses=1]
  %164 = load float* %163                         ; <float> [#uses=1]
  %165 = insertelement <4 x float> undef, float %164, i32 0 ; <<4 x float>> [#uses=1]
  %166 = shufflevector <4 x float> %165, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %167 = load <4 x float>* %14                    ; <<4 x float>> [#uses=1]
  %168 = fmul <4 x float> %119, %166              ; <<4 x float>> [#uses=1]
  %169 = fadd <4 x float> %168, %167              ; <<4 x float>> [#uses=1]
  %170 = getelementptr float* %context.vs_constants, i32 6 ; <float*> [#uses=1]
  %171 = load float* %170                         ; <float> [#uses=1]
  %172 = insertelement <4 x float> undef, float %171, i32 0 ; <<4 x float>> [#uses=1]
  %173 = shufflevector <4 x float> %172, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %174 = load <4 x float>* %13                    ; <<4 x float>> [#uses=1]
  %175 = fmul <4 x float> %119, %173              ; <<4 x float>> [#uses=1]
  %176 = fadd <4 x float> %175, %174              ; <<4 x float>> [#uses=1]
  %177 = getelementptr float* %context.vs_constants, i32 7 ; <float*> [#uses=1]
  %178 = load float* %177                         ; <float> [#uses=1]
  %179 = insertelement <4 x float> undef, float %178, i32 0 ; <<4 x float>> [#uses=1]
  %180 = shufflevector <4 x float> %179, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %181 = load <4 x float>* %12                    ; <<4 x float>> [#uses=1]
  %182 = fmul <4 x float> %119, %180              ; <<4 x float>> [#uses=1]
  %183 = fadd <4 x float> %182, %181              ; <<4 x float>> [#uses=1]
  store <4 x float> %162, <4 x float>* %15
  store <4 x float> %169, <4 x float>* %14
  store <4 x float> %176, <4 x float>* %13
  store <4 x float> %183, <4 x float>* %12
  %184 = getelementptr float* %context.vs_constants, i32 8 ; <float*> [#uses=1]
  %185 = load float* %184                         ; <float> [#uses=1]
  %186 = insertelement <4 x float> undef, float %185, i32 0 ; <<4 x float>> [#uses=1]
  %187 = shufflevector <4 x float> %186, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %188 = load <4 x float>* %15                    ; <<4 x float>> [#uses=1]
  %189 = fmul <4 x float> %127, %187              ; <<4 x float>> [#uses=1]
  %190 = fadd <4 x float> %189, %188              ; <<4 x float>> [#uses=1]
  %191 = getelementptr float* %context.vs_constants, i32 9 ; <float*> [#uses=1]
  %192 = load float* %191                         ; <float> [#uses=1]
  %193 = insertelement <4 x float> undef, float %192, i32 0 ; <<4 x float>> [#uses=1]
  %194 = shufflevector <4 x float> %193, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %195 = load <4 x float>* %14                    ; <<4 x float>> [#uses=1]
  %196 = fmul <4 x float> %127, %194              ; <<4 x float>> [#uses=1]
  %197 = fadd <4 x float> %196, %195              ; <<4 x float>> [#uses=1]
  %198 = getelementptr float* %context.vs_constants, i32 10 ; <float*> [#uses=1]
  %199 = load float* %198                         ; <float> [#uses=1]
  %200 = insertelement <4 x float> undef, float %199, i32 0 ; <<4 x float>> [#uses=1]
  %201 = shufflevector <4 x float> %200, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %202 = load <4 x float>* %13                    ; <<4 x float>> [#uses=1]
  %203 = fmul <4 x float> %127, %201              ; <<4 x float>> [#uses=1]
  %204 = fadd <4 x float> %203, %202              ; <<4 x float>> [#uses=1]
  %205 = getelementptr float* %context.vs_constants, i32 11 ; <float*> [#uses=1]
  %206 = load float* %205                         ; <float> [#uses=1]
  %207 = insertelement <4 x float> undef, float %206, i32 0 ; <<4 x float>> [#uses=1]
  %208 = shufflevector <4 x float> %207, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %209 = load <4 x float>* %12                    ; <<4 x float>> [#uses=1]
  %210 = fmul <4 x float> %127, %208              ; <<4 x float>> [#uses=1]
  %211 = fadd <4 x float> %210, %209              ; <<4 x float>> [#uses=1]
  store <4 x float> %190, <4 x float>* %15
  store <4 x float> %197, <4 x float>* %14
  store <4 x float> %204, <4 x float>* %13
  store <4 x float> %211, <4 x float>* %12
  %212 = getelementptr float* %context.vs_constants, i32 12 ; <float*> [#uses=1]
  %213 = load float* %212                         ; <float> [#uses=1]
  %214 = insertelement <4 x float> undef, float %213, i32 0 ; <<4 x float>> [#uses=1]
  %215 = shufflevector <4 x float> %214, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %216 = load <4 x float>* %15                    ; <<4 x float>> [#uses=1]
  %217 = fmul <4 x float> %135, %215              ; <<4 x float>> [#uses=1]
  %218 = fadd <4 x float> %217, %216              ; <<4 x float>> [#uses=1]
  %219 = getelementptr float* %context.vs_constants, i32 13 ; <float*> [#uses=1]
  %220 = load float* %219                         ; <float> [#uses=1]
  %221 = insertelement <4 x float> undef, float %220, i32 0 ; <<4 x float>> [#uses=1]
  %222 = shufflevector <4 x float> %221, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %223 = load <4 x float>* %14                    ; <<4 x float>> [#uses=1]
  %224 = fmul <4 x float> %135, %222              ; <<4 x float>> [#uses=1]
  %225 = fadd <4 x float> %224, %223              ; <<4 x float>> [#uses=1]
  %226 = getelementptr float* %context.vs_constants, i32 14 ; <float*> [#uses=1]
  %227 = load float* %226                         ; <float> [#uses=1]
  %228 = insertelement <4 x float> undef, float %227, i32 0 ; <<4 x float>> [#uses=1]
  %229 = shufflevector <4 x float> %228, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %230 = load <4 x float>* %13                    ; <<4 x float>> [#uses=1]
  %231 = fmul <4 x float> %135, %229              ; <<4 x float>> [#uses=1]
  %232 = fadd <4 x float> %231, %230              ; <<4 x float>> [#uses=1]
  %233 = getelementptr float* %context.vs_constants, i32 15 ; <float*> [#uses=1]
  %234 = load float* %233                         ; <float> [#uses=1]
  %235 = insertelement <4 x float> undef, float %234, i32 0 ; <<4 x float>> [#uses=1]
  %236 = shufflevector <4 x float> %235, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %237 = load <4 x float>* %12                    ; <<4 x float>> [#uses=1]
  %238 = fmul <4 x float> %135, %236              ; <<4 x float>> [#uses=1]
  %239 = fadd <4 x float> %238, %237              ; <<4 x float>> [#uses=1]
  store <4 x float> %218, <4 x float>* %23
  store <4 x float> %225, <4 x float>* %22
  store <4 x float> %232, <4 x float>* %21
  store <4 x float> %239, <4 x float>* %20
  %240 = getelementptr float* %context.vs_constants, i32 16 ; <float*> [#uses=1]
  %241 = load float* %240                         ; <float> [#uses=1]
  %242 = insertelement <4 x float> undef, float %241, i32 0 ; <<4 x float>> [#uses=1]
  %243 = shufflevector <4 x float> %242, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %244 = getelementptr float* %context.vs_constants, i32 16 ; <float*> [#uses=1]
  %245 = load float* %244                         ; <float> [#uses=1]
  %246 = insertelement <4 x float> undef, float %245, i32 0 ; <<4 x float>> [#uses=1]
  %247 = shufflevector <4 x float> %246, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %248 = fmul <4 x float> %243, %247              ; <<4 x float>> [#uses=1]
  %249 = getelementptr float* %context.vs_constants, i32 17 ; <float*> [#uses=1]
  %250 = load float* %249                         ; <float> [#uses=1]
  %251 = insertelement <4 x float> undef, float %250, i32 0 ; <<4 x float>> [#uses=1]
  %252 = shufflevector <4 x float> %251, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %253 = getelementptr float* %context.vs_constants, i32 17 ; <float*> [#uses=1]
  %254 = load float* %253                         ; <float> [#uses=1]
  %255 = insertelement <4 x float> undef, float %254, i32 0 ; <<4 x float>> [#uses=1]
  %256 = shufflevector <4 x float> %255, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %257 = fmul <4 x float> %252, %256              ; <<4 x float>> [#uses=1]
  %258 = fadd <4 x float> %248, %257              ; <<4 x float>> [#uses=1]
  %259 = getelementptr float* %context.vs_constants, i32 18 ; <float*> [#uses=1]
  %260 = load float* %259                         ; <float> [#uses=1]
  %261 = insertelement <4 x float> undef, float %260, i32 0 ; <<4 x float>> [#uses=1]
  %262 = shufflevector <4 x float> %261, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %263 = getelementptr float* %context.vs_constants, i32 18 ; <float*> [#uses=1]
  %264 = load float* %263                         ; <float> [#uses=1]
  %265 = insertelement <4 x float> undef, float %264, i32 0 ; <<4 x float>> [#uses=1]
  %266 = shufflevector <4 x float> %265, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %267 = fmul <4 x float> %262, %266              ; <<4 x float>> [#uses=1]
  %268 = fadd <4 x float> %258, %267              ; <<4 x float>> [#uses=1]
  store <4 x float> %268, <4 x float>* %11
  %269 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %270 = bitcast <4 x float> %269 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %271 = and <4 x i32> %270, <i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647> ; <<4 x i32>> [#uses=1]
  %272 = bitcast <4 x i32> %271 to <4 x float>    ; <<4 x float>> [#uses=1]
  %273 = call <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float> %272) ; <<4 x float>> [#uses=1]
  store <4 x float> %273, <4 x float>* %11
  %274 = getelementptr float* %context.vs_constants, i32 16 ; <float*> [#uses=1]
  %275 = load float* %274                         ; <float> [#uses=1]
  %276 = insertelement <4 x float> undef, float %275, i32 0 ; <<4 x float>> [#uses=1]
  %277 = shufflevector <4 x float> %276, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %278 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %279 = fmul <4 x float> %277, %278              ; <<4 x float>> [#uses=1]
  %280 = getelementptr float* %context.vs_constants, i32 17 ; <float*> [#uses=1]
  %281 = load float* %280                         ; <float> [#uses=1]
  %282 = insertelement <4 x float> undef, float %281, i32 0 ; <<4 x float>> [#uses=1]
  %283 = shufflevector <4 x float> %282, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %284 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %285 = fmul <4 x float> %283, %284              ; <<4 x float>> [#uses=1]
  %286 = getelementptr float* %context.vs_constants, i32 18 ; <float*> [#uses=1]
  %287 = load float* %286                         ; <float> [#uses=1]
  %288 = insertelement <4 x float> undef, float %287, i32 0 ; <<4 x float>> [#uses=1]
  %289 = shufflevector <4 x float> %288, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %290 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %291 = fmul <4 x float> %289, %290              ; <<4 x float>> [#uses=1]
  %292 = getelementptr float* %context.vs_constants, i32 19 ; <float*> [#uses=1]
  %293 = load float* %292                         ; <float> [#uses=1]
  %294 = insertelement <4 x float> undef, float %293, i32 0 ; <<4 x float>> [#uses=1]
  %295 = shufflevector <4 x float> %294, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %296 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %297 = fmul <4 x float> %295, %296              ; <<4 x float>> [#uses=1]
  store <4 x float> %279, <4 x float>* %15
  store <4 x float> %285, <4 x float>* %14
  store <4 x float> %291, <4 x float>* %13
  store <4 x float> %297, <4 x float>* %12
  %298 = getelementptr float* %context.vs_constants, i32 20 ; <float*> [#uses=1]
  %299 = load float* %298                         ; <float> [#uses=1]
  %300 = insertelement <4 x float> undef, float %299, i32 0 ; <<4 x float>> [#uses=1]
  %301 = shufflevector <4 x float> %300, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %302 = getelementptr float* %context.vs_constants, i32 21 ; <float*> [#uses=1]
  %303 = load float* %302                         ; <float> [#uses=1]
  %304 = insertelement <4 x float> undef, float %303, i32 0 ; <<4 x float>> [#uses=1]
  %305 = shufflevector <4 x float> %304, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %306 = getelementptr float* %context.vs_constants, i32 22 ; <float*> [#uses=1]
  %307 = load float* %306                         ; <float> [#uses=1]
  %308 = insertelement <4 x float> undef, float %307, i32 0 ; <<4 x float>> [#uses=1]
  %309 = shufflevector <4 x float> %308, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %310 = getelementptr float* %context.vs_constants, i32 23 ; <float*> [#uses=1]
  %311 = load float* %310                         ; <float> [#uses=1]
  %312 = insertelement <4 x float> undef, float %311, i32 0 ; <<4 x float>> [#uses=1]
  %313 = shufflevector <4 x float> %312, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  store <4 x float> %301, <4 x float>* %7
  store <4 x float> %305, <4 x float>* %6
  store <4 x float> %309, <4 x float>* %5
  store <4 x float> %313, <4 x float>* %4
  %314 = load <4 x float>* %7                     ; <<4 x float>> [#uses=1]
  %315 = load <4 x float>* %6                     ; <<4 x float>> [#uses=1]
  %316 = load <4 x float>* %5                     ; <<4 x float>> [#uses=1]
  %317 = load <4 x float>* %4                     ; <<4 x float>> [#uses=1]
  store <4 x float> %314, <4 x float>* %19
  store <4 x float> %315, <4 x float>* %18
  store <4 x float> %316, <4 x float>* %17
  store <4 x float> %317, <4 x float>* %16
  %318 = load <4 x float>* %15                    ; <<4 x float>> [#uses=1]
  %319 = getelementptr float* %context.vs_constants, i32 24 ; <float*> [#uses=1]
  %320 = load float* %319                         ; <float> [#uses=1]
  %321 = insertelement <4 x float> undef, float %320, i32 0 ; <<4 x float>> [#uses=1]
  %322 = shufflevector <4 x float> %321, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %323 = fmul <4 x float> %318, %322              ; <<4 x float>> [#uses=1]
  %324 = load <4 x float>* %14                    ; <<4 x float>> [#uses=1]
  %325 = getelementptr float* %context.vs_constants, i32 25 ; <float*> [#uses=1]
  %326 = load float* %325                         ; <float> [#uses=1]
  %327 = insertelement <4 x float> undef, float %326, i32 0 ; <<4 x float>> [#uses=1]
  %328 = shufflevector <4 x float> %327, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %329 = fmul <4 x float> %324, %328              ; <<4 x float>> [#uses=1]
  %330 = fadd <4 x float> %323, %329              ; <<4 x float>> [#uses=1]
  %331 = load <4 x float>* %13                    ; <<4 x float>> [#uses=1]
  %332 = getelementptr float* %context.vs_constants, i32 26 ; <float*> [#uses=1]
  %333 = load float* %332                         ; <float> [#uses=1]
  %334 = insertelement <4 x float> undef, float %333, i32 0 ; <<4 x float>> [#uses=1]
  %335 = shufflevector <4 x float> %334, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %336 = fmul <4 x float> %331, %335              ; <<4 x float>> [#uses=1]
  %337 = fadd <4 x float> %330, %336              ; <<4 x float>> [#uses=4]
  store <4 x float> %337, <4 x float>* %3
  store <4 x float> %337, <4 x float>* %2
  store <4 x float> %337, <4 x float>* %1
  store <4 x float> %337, <4 x float>* %0
  %338 = load <4 x float>* %3                     ; <<4 x float>> [#uses=1]
  %339 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> zeroinitializer, <4 x float> %338) ; <<4 x float>> [#uses=1]
  %340 = load <4 x float>* %2                     ; <<4 x float>> [#uses=1]
  %341 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> zeroinitializer, <4 x float> %340) ; <<4 x float>> [#uses=1]
  %342 = load <4 x float>* %1                     ; <<4 x float>> [#uses=1]
  %343 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> zeroinitializer, <4 x float> %342) ; <<4 x float>> [#uses=1]
  %344 = load <4 x float>* %0                     ; <<4 x float>> [#uses=1]
  %345 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, <4 x float> %344) ; <<4 x float>> [#uses=1]
  store <4 x float> %339, <4 x float>* %11
  store <4 x float> %341, <4 x float>* %10
  store <4 x float> %343, <4 x float>* %9
  store <4 x float> %345, <4 x float>* %8
  %346 = load <4 x float>* %1                     ; <<4 x float>> [#uses=1]
  %347 = fcmp ult <4 x float> zeroinitializer, %346 ; <<4 x i1>> [#uses=1]
  %348 = sext <4 x i1> %347 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %349 = and <4 x i32> <i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216>, %348 ; <<4 x i32>> [#uses=1]
  %350 = xor <4 x i32> %348, <i32 -1, i32 -1, i32 -1, i32 -1> ; <<4 x i32>> [#uses=1]
  %351 = and <4 x i32> zeroinitializer, %350      ; <<4 x i32>> [#uses=1]
  %352 = or <4 x i32> %349, %351                  ; <<4 x i32>> [#uses=1]
  %353 = bitcast <4 x i32> %352 to <4 x float>    ; <<4 x float>> [#uses=1]
  store <4 x float> %353, <4 x float>* %9
  %354 = getelementptr float* %context.vs_constants, i32 28 ; <float*> [#uses=1]
  %355 = load float* %354                         ; <float> [#uses=1]
  %356 = insertelement <4 x float> undef, float %355, i32 0 ; <<4 x float>> [#uses=1]
  %357 = shufflevector <4 x float> %356, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %358 = load <4 x float>* %7                     ; <<4 x float>> [#uses=1]
  %359 = fadd <4 x float> %357, %358              ; <<4 x float>> [#uses=1]
  %360 = getelementptr float* %context.vs_constants, i32 29 ; <float*> [#uses=1]
  %361 = load float* %360                         ; <float> [#uses=1]
  %362 = insertelement <4 x float> undef, float %361, i32 0 ; <<4 x float>> [#uses=1]
  %363 = shufflevector <4 x float> %362, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %364 = load <4 x float>* %6                     ; <<4 x float>> [#uses=1]
  %365 = fadd <4 x float> %363, %364              ; <<4 x float>> [#uses=1]
  %366 = getelementptr float* %context.vs_constants, i32 30 ; <float*> [#uses=1]
  %367 = load float* %366                         ; <float> [#uses=1]
  %368 = insertelement <4 x float> undef, float %367, i32 0 ; <<4 x float>> [#uses=1]
  %369 = shufflevector <4 x float> %368, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %370 = load <4 x float>* %5                     ; <<4 x float>> [#uses=1]
  %371 = fadd <4 x float> %369, %370              ; <<4 x float>> [#uses=1]
  %372 = getelementptr float* %context.vs_constants, i32 31 ; <float*> [#uses=1]
  %373 = load float* %372                         ; <float> [#uses=1]
  %374 = insertelement <4 x float> undef, float %373, i32 0 ; <<4 x float>> [#uses=1]
  %375 = shufflevector <4 x float> %374, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %376 = load <4 x float>* %4                     ; <<4 x float>> [#uses=1]
  %377 = fadd <4 x float> %375, %376              ; <<4 x float>> [#uses=1]
  store <4 x float> %359, <4 x float>* %7
  store <4 x float> %365, <4 x float>* %6
  store <4 x float> %371, <4 x float>* %5
  store <4 x float> %377, <4 x float>* %4
  %378 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %379 = getelementptr float* %context.vs_constants, i32 32 ; <float*> [#uses=1]
  %380 = load float* %379                         ; <float> [#uses=1]
  %381 = insertelement <4 x float> undef, float %380, i32 0 ; <<4 x float>> [#uses=1]
  %382 = shufflevector <4 x float> %381, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %383 = load <4 x float>* %7                     ; <<4 x float>> [#uses=1]
  %384 = fmul <4 x float> %378, %382              ; <<4 x float>> [#uses=1]
  %385 = fadd <4 x float> %384, %383              ; <<4 x float>> [#uses=1]
  %386 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %387 = getelementptr float* %context.vs_constants, i32 33 ; <float*> [#uses=1]
  %388 = load float* %387                         ; <float> [#uses=1]
  %389 = insertelement <4 x float> undef, float %388, i32 0 ; <<4 x float>> [#uses=1]
  %390 = shufflevector <4 x float> %389, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %391 = load <4 x float>* %6                     ; <<4 x float>> [#uses=1]
  %392 = fmul <4 x float> %386, %390              ; <<4 x float>> [#uses=1]
  %393 = fadd <4 x float> %392, %391              ; <<4 x float>> [#uses=1]
  %394 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %395 = getelementptr float* %context.vs_constants, i32 34 ; <float*> [#uses=1]
  %396 = load float* %395                         ; <float> [#uses=1]
  %397 = insertelement <4 x float> undef, float %396, i32 0 ; <<4 x float>> [#uses=1]
  %398 = shufflevector <4 x float> %397, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %399 = load <4 x float>* %5                     ; <<4 x float>> [#uses=1]
  %400 = fmul <4 x float> %394, %398              ; <<4 x float>> [#uses=1]
  %401 = fadd <4 x float> %400, %399              ; <<4 x float>> [#uses=1]
  %402 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %403 = getelementptr float* %context.vs_constants, i32 35 ; <float*> [#uses=1]
  %404 = load float* %403                         ; <float> [#uses=1]
  %405 = insertelement <4 x float> undef, float %404, i32 0 ; <<4 x float>> [#uses=1]
  %406 = shufflevector <4 x float> %405, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %407 = load <4 x float>* %4                     ; <<4 x float>> [#uses=1]
  %408 = fmul <4 x float> %402, %406              ; <<4 x float>> [#uses=1]
  %409 = fadd <4 x float> %408, %407              ; <<4 x float>> [#uses=1]
  store <4 x float> %385, <4 x float>* %7
  store <4 x float> %393, <4 x float>* %6
  store <4 x float> %401, <4 x float>* %5
  store <4 x float> %409, <4 x float>* %4
  %410 = load <4 x float>* %9                     ; <<4 x float>> [#uses=1]
  %411 = getelementptr float* %context.vs_constants, i32 36 ; <float*> [#uses=1]
  %412 = load float* %411                         ; <float> [#uses=1]
  %413 = insertelement <4 x float> undef, float %412, i32 0 ; <<4 x float>> [#uses=1]
  %414 = shufflevector <4 x float> %413, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %415 = load <4 x float>* %7                     ; <<4 x float>> [#uses=1]
  %416 = fmul <4 x float> %410, %414              ; <<4 x float>> [#uses=1]
  %417 = fadd <4 x float> %416, %415              ; <<4 x float>> [#uses=1]
  %418 = load <4 x float>* %9                     ; <<4 x float>> [#uses=1]
  %419 = getelementptr float* %context.vs_constants, i32 37 ; <float*> [#uses=1]
  %420 = load float* %419                         ; <float> [#uses=1]
  %421 = insertelement <4 x float> undef, float %420, i32 0 ; <<4 x float>> [#uses=1]
  %422 = shufflevector <4 x float> %421, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %423 = load <4 x float>* %6                     ; <<4 x float>> [#uses=1]
  %424 = fmul <4 x float> %418, %422              ; <<4 x float>> [#uses=1]
  %425 = fadd <4 x float> %424, %423              ; <<4 x float>> [#uses=1]
  %426 = load <4 x float>* %9                     ; <<4 x float>> [#uses=1]
  %427 = getelementptr float* %context.vs_constants, i32 38 ; <float*> [#uses=1]
  %428 = load float* %427                         ; <float> [#uses=1]
  %429 = insertelement <4 x float> undef, float %428, i32 0 ; <<4 x float>> [#uses=1]
  %430 = shufflevector <4 x float> %429, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %431 = load <4 x float>* %5                     ; <<4 x float>> [#uses=1]
  %432 = fmul <4 x float> %426, %430              ; <<4 x float>> [#uses=1]
  %433 = fadd <4 x float> %432, %431              ; <<4 x float>> [#uses=1]
  store <4 x float> %417, <4 x float>* %19
  store <4 x float> %425, <4 x float>* %18
  store <4 x float> %433, <4 x float>* %17
  %output0.x = load <4 x float>* %23              ; <<4 x float>> [#uses=4]
  %output0.y = load <4 x float>* %22              ; <<4 x float>> [#uses=4]
  %output0.z = load <4 x float>* %21              ; <<4 x float>> [#uses=4]
  %output0.w = load <4 x float>* %20              ; <<4 x float>> [#uses=4]
  %434 = extractelement <4 x float> %output0.x, i32 0 ; <float> [#uses=1]
  %435 = insertelement <4 x float> zeroinitializer, float %434, i32 0 ; <<4 x float>> [#uses=1]
  %436 = extractelement <4 x float> %output0.x, i32 1 ; <float> [#uses=1]
  %437 = insertelement <4 x float> zeroinitializer, float %436, i32 0 ; <<4 x float>> [#uses=1]
  %438 = extractelement <4 x float> %output0.x, i32 2 ; <float> [#uses=1]
  %439 = insertelement <4 x float> zeroinitializer, float %438, i32 0 ; <<4 x float>> [#uses=1]
  %440 = extractelement <4 x float> %output0.x, i32 3 ; <float> [#uses=1]
  %441 = insertelement <4 x float> zeroinitializer, float %440, i32 0 ; <<4 x float>> [#uses=1]
  %442 = extractelement <4 x float> %output0.y, i32 0 ; <float> [#uses=1]
  %443 = insertelement <4 x float> %435, float %442, i32 1 ; <<4 x float>> [#uses=1]
  %444 = extractelement <4 x float> %output0.y, i32 1 ; <float> [#uses=1]
  %445 = insertelement <4 x float> %437, float %444, i32 1 ; <<4 x float>> [#uses=1]
  %446 = extractelement <4 x float> %output0.y, i32 2 ; <float> [#uses=1]
  %447 = insertelement <4 x float> %439, float %446, i32 1 ; <<4 x float>> [#uses=1]
  %448 = extractelement <4 x float> %output0.y, i32 3 ; <float> [#uses=1]
  %449 = insertelement <4 x float> %441, float %448, i32 1 ; <<4 x float>> [#uses=1]
  %450 = extractelement <4 x float> %output0.z, i32 0 ; <float> [#uses=1]
  %451 = insertelement <4 x float> %443, float %450, i32 2 ; <<4 x float>> [#uses=1]
  %452 = extractelement <4 x float> %output0.z, i32 1 ; <float> [#uses=1]
  %453 = insertelement <4 x float> %445, float %452, i32 2 ; <<4 x float>> [#uses=1]
  %454 = extractelement <4 x float> %output0.z, i32 2 ; <float> [#uses=1]
  %455 = insertelement <4 x float> %447, float %454, i32 2 ; <<4 x float>> [#uses=1]
  %456 = extractelement <4 x float> %output0.z, i32 3 ; <float> [#uses=1]
  %457 = insertelement <4 x float> %449, float %456, i32 2 ; <<4 x float>> [#uses=1]
  %458 = extractelement <4 x float> %output0.w, i32 0 ; <float> [#uses=1]
  %459 = insertelement <4 x float> %451, float %458, i32 3 ; <<4 x float>> [#uses=4]
  %460 = extractelement <4 x float> %output0.w, i32 1 ; <float> [#uses=1]
  %461 = insertelement <4 x float> %453, float %460, i32 3 ; <<4 x float>> [#uses=4]
  %462 = extractelement <4 x float> %output0.w, i32 2 ; <float> [#uses=1]
  %463 = insertelement <4 x float> %455, float %462, i32 3 ; <<4 x float>> [#uses=4]
  %464 = extractelement <4 x float> %output0.w, i32 3 ; <float> [#uses=1]
  %465 = insertelement <4 x float> %457, float %464, i32 3 ; <<4 x float>> [#uses=4]
  %466 = getelementptr %vertex_header3* %27, i32 0 ; <%vertex_header3*> [#uses=2]
  %467 = getelementptr %vertex_header3* %27, i32 1 ; <%vertex_header3*> [#uses=2]
  %468 = getelementptr %vertex_header3* %27, i32 2 ; <%vertex_header3*> [#uses=2]
  %469 = getelementptr %vertex_header3* %27, i32 3 ; <%vertex_header3*> [#uses=2]
  %.id_ptr = getelementptr %vertex_header3* %466, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr = getelementptr %vertex_header3* %466, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr
  %470 = getelementptr [3 x [4 x float]]* %.data_ptr, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %471 = extractelement <4 x float> %459, i32 0   ; <float> [#uses=1]
  %472 = extractelement <4 x float> %459, i32 1   ; <float> [#uses=1]
  %473 = extractelement <4 x float> %459, i32 2   ; <float> [#uses=1]
  %474 = extractelement <4 x float> %459, i32 3   ; <float> [#uses=1]
  %475 = getelementptr float* %470, i32 0         ; <float*> [#uses=1]
  %476 = getelementptr float* %470, i32 1         ; <float*> [#uses=1]
  %477 = getelementptr float* %470, i32 2         ; <float*> [#uses=1]
  %478 = getelementptr float* %470, i32 3         ; <float*> [#uses=1]
  store float %471, float* %475
  store float %472, float* %476
  store float %473, float* %477
  store float %474, float* %478
  %.id_ptr33 = getelementptr %vertex_header3* %467, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr34 = getelementptr %vertex_header3* %467, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr33
  %479 = getelementptr [3 x [4 x float]]* %.data_ptr34, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %480 = extractelement <4 x float> %461, i32 0   ; <float> [#uses=1]
  %481 = extractelement <4 x float> %461, i32 1   ; <float> [#uses=1]
  %482 = extractelement <4 x float> %461, i32 2   ; <float> [#uses=1]
  %483 = extractelement <4 x float> %461, i32 3   ; <float> [#uses=1]
  %484 = getelementptr float* %479, i32 0         ; <float*> [#uses=1]
  %485 = getelementptr float* %479, i32 1         ; <float*> [#uses=1]
  %486 = getelementptr float* %479, i32 2         ; <float*> [#uses=1]
  %487 = getelementptr float* %479, i32 3         ; <float*> [#uses=1]
  store float %480, float* %484
  store float %481, float* %485
  store float %482, float* %486
  store float %483, float* %487
  %.id_ptr35 = getelementptr %vertex_header3* %468, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr36 = getelementptr %vertex_header3* %468, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr35
  %488 = getelementptr [3 x [4 x float]]* %.data_ptr36, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %489 = extractelement <4 x float> %463, i32 0   ; <float> [#uses=1]
  %490 = extractelement <4 x float> %463, i32 1   ; <float> [#uses=1]
  %491 = extractelement <4 x float> %463, i32 2   ; <float> [#uses=1]
  %492 = extractelement <4 x float> %463, i32 3   ; <float> [#uses=1]
  %493 = getelementptr float* %488, i32 0         ; <float*> [#uses=1]
  %494 = getelementptr float* %488, i32 1         ; <float*> [#uses=1]
  %495 = getelementptr float* %488, i32 2         ; <float*> [#uses=1]
  %496 = getelementptr float* %488, i32 3         ; <float*> [#uses=1]
  store float %489, float* %493
  store float %490, float* %494
  store float %491, float* %495
  store float %492, float* %496
  %.id_ptr37 = getelementptr %vertex_header3* %469, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr38 = getelementptr %vertex_header3* %469, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr37
  %497 = getelementptr [3 x [4 x float]]* %.data_ptr38, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %498 = extractelement <4 x float> %465, i32 0   ; <float> [#uses=1]
  %499 = extractelement <4 x float> %465, i32 1   ; <float> [#uses=1]
  %500 = extractelement <4 x float> %465, i32 2   ; <float> [#uses=1]
  %501 = extractelement <4 x float> %465, i32 3   ; <float> [#uses=1]
  %502 = getelementptr float* %497, i32 0         ; <float*> [#uses=1]
  %503 = getelementptr float* %497, i32 1         ; <float*> [#uses=1]
  %504 = getelementptr float* %497, i32 2         ; <float*> [#uses=1]
  %505 = getelementptr float* %497, i32 3         ; <float*> [#uses=1]
  store float %498, float* %502
  store float %499, float* %503
  store float %500, float* %504
  store float %501, float* %505
  %output1.x = load <4 x float>* %19              ; <<4 x float>> [#uses=4]
  %output1.y = load <4 x float>* %18              ; <<4 x float>> [#uses=4]
  %output1.z = load <4 x float>* %17              ; <<4 x float>> [#uses=4]
  %output1.w = load <4 x float>* %16              ; <<4 x float>> [#uses=4]
  %506 = extractelement <4 x float> %output1.x, i32 0 ; <float> [#uses=1]
  %507 = insertelement <4 x float> zeroinitializer, float %506, i32 0 ; <<4 x float>> [#uses=1]
  %508 = extractelement <4 x float> %output1.x, i32 1 ; <float> [#uses=1]
  %509 = insertelement <4 x float> zeroinitializer, float %508, i32 0 ; <<4 x float>> [#uses=1]
  %510 = extractelement <4 x float> %output1.x, i32 2 ; <float> [#uses=1]
  %511 = insertelement <4 x float> zeroinitializer, float %510, i32 0 ; <<4 x float>> [#uses=1]
  %512 = extractelement <4 x float> %output1.x, i32 3 ; <float> [#uses=1]
  %513 = insertelement <4 x float> zeroinitializer, float %512, i32 0 ; <<4 x float>> [#uses=1]
  %514 = extractelement <4 x float> %output1.y, i32 0 ; <float> [#uses=1]
  %515 = insertelement <4 x float> %507, float %514, i32 1 ; <<4 x float>> [#uses=1]
  %516 = extractelement <4 x float> %output1.y, i32 1 ; <float> [#uses=1]
  %517 = insertelement <4 x float> %509, float %516, i32 1 ; <<4 x float>> [#uses=1]
  %518 = extractelement <4 x float> %output1.y, i32 2 ; <float> [#uses=1]
  %519 = insertelement <4 x float> %511, float %518, i32 1 ; <<4 x float>> [#uses=1]
  %520 = extractelement <4 x float> %output1.y, i32 3 ; <float> [#uses=1]
  %521 = insertelement <4 x float> %513, float %520, i32 1 ; <<4 x float>> [#uses=1]
  %522 = extractelement <4 x float> %output1.z, i32 0 ; <float> [#uses=1]
  %523 = insertelement <4 x float> %515, float %522, i32 2 ; <<4 x float>> [#uses=1]
  %524 = extractelement <4 x float> %output1.z, i32 1 ; <float> [#uses=1]
  %525 = insertelement <4 x float> %517, float %524, i32 2 ; <<4 x float>> [#uses=1]
  %526 = extractelement <4 x float> %output1.z, i32 2 ; <float> [#uses=1]
  %527 = insertelement <4 x float> %519, float %526, i32 2 ; <<4 x float>> [#uses=1]
  %528 = extractelement <4 x float> %output1.z, i32 3 ; <float> [#uses=1]
  %529 = insertelement <4 x float> %521, float %528, i32 2 ; <<4 x float>> [#uses=1]
  %530 = extractelement <4 x float> %output1.w, i32 0 ; <float> [#uses=1]
  %531 = insertelement <4 x float> %523, float %530, i32 3 ; <<4 x float>> [#uses=4]
  %532 = extractelement <4 x float> %output1.w, i32 1 ; <float> [#uses=1]
  %533 = insertelement <4 x float> %525, float %532, i32 3 ; <<4 x float>> [#uses=4]
  %534 = extractelement <4 x float> %output1.w, i32 2 ; <float> [#uses=1]
  %535 = insertelement <4 x float> %527, float %534, i32 3 ; <<4 x float>> [#uses=4]
  %536 = extractelement <4 x float> %output1.w, i32 3 ; <float> [#uses=1]
  %537 = insertelement <4 x float> %529, float %536, i32 3 ; <<4 x float>> [#uses=4]
  %538 = getelementptr %vertex_header3* %27, i32 0 ; <%vertex_header3*> [#uses=2]
  %539 = getelementptr %vertex_header3* %27, i32 1 ; <%vertex_header3*> [#uses=2]
  %540 = getelementptr %vertex_header3* %27, i32 2 ; <%vertex_header3*> [#uses=2]
  %541 = getelementptr %vertex_header3* %27, i32 3 ; <%vertex_header3*> [#uses=2]
  %.id_ptr39 = getelementptr %vertex_header3* %538, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr40 = getelementptr %vertex_header3* %538, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr39
  %542 = getelementptr [3 x [4 x float]]* %.data_ptr40, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %543 = extractelement <4 x float> %531, i32 0   ; <float> [#uses=1]
  %544 = extractelement <4 x float> %531, i32 1   ; <float> [#uses=1]
  %545 = extractelement <4 x float> %531, i32 2   ; <float> [#uses=1]
  %546 = extractelement <4 x float> %531, i32 3   ; <float> [#uses=1]
  %547 = getelementptr float* %542, i32 0         ; <float*> [#uses=1]
  %548 = getelementptr float* %542, i32 1         ; <float*> [#uses=1]
  %549 = getelementptr float* %542, i32 2         ; <float*> [#uses=1]
  %550 = getelementptr float* %542, i32 3         ; <float*> [#uses=1]
  store float %543, float* %547
  store float %544, float* %548
  store float %545, float* %549
  store float %546, float* %550
  %.id_ptr41 = getelementptr %vertex_header3* %539, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr42 = getelementptr %vertex_header3* %539, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr41
  %551 = getelementptr [3 x [4 x float]]* %.data_ptr42, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %552 = extractelement <4 x float> %533, i32 0   ; <float> [#uses=1]
  %553 = extractelement <4 x float> %533, i32 1   ; <float> [#uses=1]
  %554 = extractelement <4 x float> %533, i32 2   ; <float> [#uses=1]
  %555 = extractelement <4 x float> %533, i32 3   ; <float> [#uses=1]
  %556 = getelementptr float* %551, i32 0         ; <float*> [#uses=1]
  %557 = getelementptr float* %551, i32 1         ; <float*> [#uses=1]
  %558 = getelementptr float* %551, i32 2         ; <float*> [#uses=1]
  %559 = getelementptr float* %551, i32 3         ; <float*> [#uses=1]
  store float %552, float* %556
  store float %553, float* %557
  store float %554, float* %558
  store float %555, float* %559
  %.id_ptr43 = getelementptr %vertex_header3* %540, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr44 = getelementptr %vertex_header3* %540, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr43
  %560 = getelementptr [3 x [4 x float]]* %.data_ptr44, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %561 = extractelement <4 x float> %535, i32 0   ; <float> [#uses=1]
  %562 = extractelement <4 x float> %535, i32 1   ; <float> [#uses=1]
  %563 = extractelement <4 x float> %535, i32 2   ; <float> [#uses=1]
  %564 = extractelement <4 x float> %535, i32 3   ; <float> [#uses=1]
  %565 = getelementptr float* %560, i32 0         ; <float*> [#uses=1]
  %566 = getelementptr float* %560, i32 1         ; <float*> [#uses=1]
  %567 = getelementptr float* %560, i32 2         ; <float*> [#uses=1]
  %568 = getelementptr float* %560, i32 3         ; <float*> [#uses=1]
  store float %561, float* %565
  store float %562, float* %566
  store float %563, float* %567
  store float %564, float* %568
  %.id_ptr45 = getelementptr %vertex_header3* %541, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr46 = getelementptr %vertex_header3* %541, i32 0, i32 2 ; <[3 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr45
  %569 = getelementptr [3 x [4 x float]]* %.data_ptr46, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %570 = extractelement <4 x float> %537, i32 0   ; <float> [#uses=1]
  %571 = extractelement <4 x float> %537, i32 1   ; <float> [#uses=1]
  %572 = extractelement <4 x float> %537, i32 2   ; <float> [#uses=1]
  %573 = extractelement <4 x float> %537, i32 3   ; <float> [#uses=1]
  %574 = getelementptr float* %569, i32 0         ; <float*> [#uses=1]
  %575 = getelementptr float* %569, i32 1         ; <float*> [#uses=1]
  %576 = getelementptr float* %569, i32 2         ; <float*> [#uses=1]
  %577 = getelementptr float* %569, i32 3         ; <float*> [#uses=1]
  store float %570, float* %574
  store float %571, float* %575
  store float %572, float* %576
  store float %573, float* %577
  %578 = add i32 %25, 4                           ; <i32> [#uses=2]
  %579 = icmp uge i32 %578, %24                   ; <i1> [#uses=1]
  br i1 %579, label %580, label %loop

; <label>:580                                     ; preds = %loop
  ret void
}

draw/draw_llvm.c:813:draw_llvm_generate: Assertion `0' failed.
debug_get_bool_option: GALLIUM_ABORT_ON_ASSERT = TRUE
LLVM ERROR: Cannot yet select: 0x3259150: i32 = fadd 0x325c280, 0x32633f0 [ORD=707] [ID=43]
  0x325c280: i32,ch = CopyFromReg 0x18a1008, 0x325b470 [ORD=707] [ID=31]
    0x18a1008: ch = EntryToken [ORD=707] [ID=0]
    0x325b470: i32 = Register %reg1049 [ORD=707] [ID=4]
  0x32633f0: i32,ch = CopyFromReg 0x18a1008, 0x3263cf0 [ORD=707] [ID=32]
    0x18a1008: ch = EntryToken [ORD=707] [ID=0]
    0x3263cf0: i32 = Register %reg1050 [ORD=707] [ID=5]