debug_get_flags_option: RADEON_DEBUG = 0x0
debug_get_bool_option: RADEON_NO_TCL = FALSE
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: GALLIUM_GALAHAD = FALSE
debug_get_flags_option: GALLIVM_DEBUG = 0x1 (tgsi)
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_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: MESA_MVP_DP4 = FALSE
debug_get_flags_option: GALLIVM_DEBUG = 0x1 (tgsi)
Integer arithmetic operators only work with integral types!
  %224 = mul <4 x float> %163, %223               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %229 = mul <4 x float> %163, %228               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %234 = mul <4 x float> %163, %233               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %239 = mul <4 x float> %163, %238               ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %224 = mul <4 x float> %163, %223               ; <<4 x float>> [#uses=1]
  store <4 x float> %224, <4 x float>* %11
Instruction does not dominate all uses!
  %229 = mul <4 x float> %163, %228               ; <<4 x float>> [#uses=1]
  store <4 x float> %229, <4 x float>* %10
Instruction does not dominate all uses!
  %234 = mul <4 x float> %163, %233               ; <<4 x float>> [#uses=1]
  store <4 x float> %234, <4 x float>* %9
Instruction does not dominate all uses!
  %239 = mul <4 x float> %163, %238               ; <<4 x float>> [#uses=1]
  store <4 x float> %239, <4 x float>* %8
Integer arithmetic operators only work with integral types!
  %245 = mul <4 x float> %171, %243               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %246 = add <4 x float> %245, %244               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %252 = mul <4 x float> %171, %250               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %253 = add <4 x float> %252, %251               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %259 = mul <4 x float> %171, %257               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %260 = add <4 x float> %259, %258               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %266 = mul <4 x float> %171, %264               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %267 = add <4 x float> %266, %265               ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %246 = add <4 x float> %245, %244               ; <<4 x float>> [#uses=1]
  store <4 x float> %246, <4 x float>* %11
Instruction does not dominate all uses!
  %253 = add <4 x float> %252, %251               ; <<4 x float>> [#uses=1]
  store <4 x float> %253, <4 x float>* %10
Instruction does not dominate all uses!
  %260 = add <4 x float> %259, %258               ; <<4 x float>> [#uses=1]
  store <4 x float> %260, <4 x float>* %9
Instruction does not dominate all uses!
  %267 = add <4 x float> %266, %265               ; <<4 x float>> [#uses=1]
  store <4 x float> %267, <4 x float>* %8
Integer arithmetic operators only work with integral types!
  %273 = mul <4 x float> %179, %271               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %274 = add <4 x float> %273, %272               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %280 = mul <4 x float> %179, %278               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %281 = add <4 x float> %280, %279               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %287 = mul <4 x float> %179, %285               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %288 = add <4 x float> %287, %286               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %294 = mul <4 x float> %179, %292               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %295 = add <4 x float> %294, %293               ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %274 = add <4 x float> %273, %272               ; <<4 x float>> [#uses=1]
  store <4 x float> %274, <4 x float>* %11
Instruction does not dominate all uses!
  %281 = add <4 x float> %280, %279               ; <<4 x float>> [#uses=1]
  store <4 x float> %281, <4 x float>* %10
Instruction does not dominate all uses!
  %288 = add <4 x float> %287, %286               ; <<4 x float>> [#uses=1]
  store <4 x float> %288, <4 x float>* %9
Instruction does not dominate all uses!
  %295 = add <4 x float> %294, %293               ; <<4 x float>> [#uses=1]
  store <4 x float> %295, <4 x float>* %8
Integer arithmetic operators only work with integral types!
  %301 = mul <4 x float> %187, %299               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %302 = add <4 x float> %301, %300               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %308 = mul <4 x float> %187, %306               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %309 = add <4 x float> %308, %307               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %315 = mul <4 x float> %187, %313               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %316 = add <4 x float> %315, %314               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %322 = mul <4 x float> %187, %320               ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %323 = add <4 x float> %322, %321               ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %302 = add <4 x float> %301, %300               ; <<4 x float>> [#uses=1]
  store <4 x float> %302, <4 x float>* %3
Instruction does not dominate all uses!
  %309 = add <4 x float> %308, %307               ; <<4 x float>> [#uses=1]
  store <4 x float> %309, <4 x float>* %2
Instruction does not dominate all uses!
  %316 = add <4 x float> %315, %314               ; <<4 x float>> [#uses=1]
  store <4 x float> %316, <4 x float>* %1
Instruction does not dominate all uses!
  %323 = add <4 x float> %322, %321               ; <<4 x float>> [#uses=1]
  store <4 x float> %323, <4 x float>* %0
Broken module found, verification continues.

define void @draw_llvm_shader(%draw_jit_context* noalias %context, %vertex_header4* 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=3]
  %1 = alloca <4 x float>                         ; <<4 x float>*> [#uses=3]
  %2 = alloca <4 x float>                         ; <<4 x float>*> [#uses=3]
  %3 = alloca <4 x float>                         ; <<4 x float>*> [#uses=3]
  %4 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %5 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %6 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %7 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %8 = alloca <4 x float>                         ; <<4 x float>*> [#uses=6]
  %9 = alloca <4 x float>                         ; <<4 x float>*> [#uses=6]
  %10 = alloca <4 x float>                        ; <<4 x float>*> [#uses=6]
  %11 = alloca <4 x float>                        ; <<4 x float>*> [#uses=6]
  %12 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %13 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %14 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %15 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %16 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %17 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %18 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %19 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %20 = add i32 %start, %count                    ; <i32> [#uses=1]
  br label %loop

loop:                                             ; preds = %loop, %entry
  %21 = phi i32 [ %start, %entry ], [ %548, %loop ] ; <i32> [#uses=6]
  %22 = sub i32 %21, %start                       ; <i32> [#uses=1]
  %23 = getelementptr %vertex_header4* %io, i32 %22 ; <%vertex_header4*> [#uses=12]
  %24 = add i32 %21, 0                            ; <i32> [#uses=4]
  %25 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %26 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr = getelementptr %pipe_vertex_buffer* %25, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride = load i32* %.stride_ptr               ; <i32> [#uses=1]
  %.max_index_ptr = getelementptr %pipe_vertex_buffer* %25, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index = load i32* %.max_index_ptr         ; <i32> [#uses=2]
  %.buffer_offset_ptr = getelementptr %pipe_vertex_buffer* %25, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset = load i32* %.buffer_offset_ptr ; <i32> [#uses=1]
  %27 = icmp ule i32 %24, %.max_index             ; <i1> [#uses=1]
  %28 = select i1 %27, i32 %24, i32 %.max_index   ; <i32> [#uses=1]
  %29 = mul i32 %.stride, %28                     ; <i32> [#uses=1]
  %vbuffer = load i8** %26                        ; <i8*> [#uses=1]
  %30 = add i32 %29, %.buffer_offset              ; <i32> [#uses=1]
  %31 = add i32 %30, 0                            ; <i32> [#uses=1]
  %32 = getelementptr i8* %vbuffer, i32 %31       ; <i8*> [#uses=2]
  %src_tmp = getelementptr i8* %32, i32 0         ; <i8*> [#uses=1]
  %33 = bitcast i8* %src_tmp to float*            ; <float*> [#uses=1]
  %34 = load float* %33                           ; <float> [#uses=1]
  %35 = insertelement <4 x float> zeroinitializer, float %34, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp1 = getelementptr i8* %32, i32 4        ; <i8*> [#uses=1]
  %36 = bitcast i8* %src_tmp1 to float*           ; <float*> [#uses=1]
  %37 = load float* %36                           ; <float> [#uses=1]
  %38 = insertelement <4 x float> %35, float %37, i32 1 ; <<4 x float>> [#uses=1]
  %39 = insertelement <4 x float> %38, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %40 = insertelement <4 x float> %39, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %41 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %42 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr2 = getelementptr %pipe_vertex_buffer* %41, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride3 = load i32* %.stride_ptr2             ; <i32> [#uses=1]
  %.max_index_ptr4 = getelementptr %pipe_vertex_buffer* %41, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index5 = load i32* %.max_index_ptr4       ; <i32> [#uses=2]
  %.buffer_offset_ptr6 = getelementptr %pipe_vertex_buffer* %41, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset7 = load i32* %.buffer_offset_ptr6 ; <i32> [#uses=1]
  %43 = icmp ule i32 %24, %.max_index5            ; <i1> [#uses=1]
  %44 = select i1 %43, i32 %24, i32 %.max_index5  ; <i32> [#uses=1]
  %45 = mul i32 %.stride3, %44                    ; <i32> [#uses=1]
  %vbuffer8 = load i8** %42                       ; <i8*> [#uses=1]
  %46 = add i32 %45, %.buffer_offset7             ; <i32> [#uses=1]
  %47 = add i32 %46, 8                            ; <i32> [#uses=1]
  %48 = getelementptr i8* %vbuffer8, i32 %47      ; <i8*> [#uses=2]
  %src_tmp9 = getelementptr i8* %48, i32 0        ; <i8*> [#uses=1]
  %49 = bitcast i8* %src_tmp9 to float*           ; <float*> [#uses=1]
  %50 = load float* %49                           ; <float> [#uses=1]
  %51 = insertelement <4 x float> zeroinitializer, float %50, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp10 = getelementptr i8* %48, i32 4       ; <i8*> [#uses=1]
  %52 = bitcast i8* %src_tmp10 to float*          ; <float*> [#uses=1]
  %53 = load float* %52                           ; <float> [#uses=1]
  %54 = insertelement <4 x float> %51, float %53, i32 1 ; <<4 x float>> [#uses=1]
  %55 = insertelement <4 x float> %54, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %56 = insertelement <4 x float> %55, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %57 = add i32 %21, 1                            ; <i32> [#uses=4]
  %58 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %59 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr11 = getelementptr %pipe_vertex_buffer* %58, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride12 = load i32* %.stride_ptr11           ; <i32> [#uses=1]
  %.max_index_ptr13 = getelementptr %pipe_vertex_buffer* %58, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index14 = load i32* %.max_index_ptr13     ; <i32> [#uses=2]
  %.buffer_offset_ptr15 = getelementptr %pipe_vertex_buffer* %58, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset16 = load i32* %.buffer_offset_ptr15 ; <i32> [#uses=1]
  %60 = icmp ule i32 %57, %.max_index14           ; <i1> [#uses=1]
  %61 = select i1 %60, i32 %57, i32 %.max_index14 ; <i32> [#uses=1]
  %62 = mul i32 %.stride12, %61                   ; <i32> [#uses=1]
  %vbuffer17 = load i8** %59                      ; <i8*> [#uses=1]
  %63 = add i32 %62, %.buffer_offset16            ; <i32> [#uses=1]
  %64 = add i32 %63, 0                            ; <i32> [#uses=1]
  %65 = getelementptr i8* %vbuffer17, i32 %64     ; <i8*> [#uses=2]
  %src_tmp18 = getelementptr i8* %65, i32 0       ; <i8*> [#uses=1]
  %66 = bitcast i8* %src_tmp18 to float*          ; <float*> [#uses=1]
  %67 = load float* %66                           ; <float> [#uses=1]
  %68 = insertelement <4 x float> zeroinitializer, float %67, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp19 = getelementptr i8* %65, i32 4       ; <i8*> [#uses=1]
  %69 = bitcast i8* %src_tmp19 to float*          ; <float*> [#uses=1]
  %70 = load float* %69                           ; <float> [#uses=1]
  %71 = insertelement <4 x float> %68, float %70, i32 1 ; <<4 x float>> [#uses=1]
  %72 = insertelement <4 x float> %71, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %73 = insertelement <4 x float> %72, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %74 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %75 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr20 = getelementptr %pipe_vertex_buffer* %74, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride21 = load i32* %.stride_ptr20           ; <i32> [#uses=1]
  %.max_index_ptr22 = getelementptr %pipe_vertex_buffer* %74, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index23 = load i32* %.max_index_ptr22     ; <i32> [#uses=2]
  %.buffer_offset_ptr24 = getelementptr %pipe_vertex_buffer* %74, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset25 = load i32* %.buffer_offset_ptr24 ; <i32> [#uses=1]
  %76 = icmp ule i32 %57, %.max_index23           ; <i1> [#uses=1]
  %77 = select i1 %76, i32 %57, i32 %.max_index23 ; <i32> [#uses=1]
  %78 = mul i32 %.stride21, %77                   ; <i32> [#uses=1]
  %vbuffer26 = load i8** %75                      ; <i8*> [#uses=1]
  %79 = add i32 %78, %.buffer_offset25            ; <i32> [#uses=1]
  %80 = add i32 %79, 8                            ; <i32> [#uses=1]
  %81 = getelementptr i8* %vbuffer26, i32 %80     ; <i8*> [#uses=2]
  %src_tmp27 = getelementptr i8* %81, i32 0       ; <i8*> [#uses=1]
  %82 = bitcast i8* %src_tmp27 to float*          ; <float*> [#uses=1]
  %83 = load float* %82                           ; <float> [#uses=1]
  %84 = insertelement <4 x float> zeroinitializer, float %83, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp28 = getelementptr i8* %81, i32 4       ; <i8*> [#uses=1]
  %85 = bitcast i8* %src_tmp28 to float*          ; <float*> [#uses=1]
  %86 = load float* %85                           ; <float> [#uses=1]
  %87 = insertelement <4 x float> %84, float %86, i32 1 ; <<4 x float>> [#uses=1]
  %88 = insertelement <4 x float> %87, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %89 = insertelement <4 x float> %88, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %90 = add i32 %21, 2                            ; <i32> [#uses=4]
  %91 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %92 = getelementptr i8** %vbuffers, i64 0       ; <i8**> [#uses=1]
  %.stride_ptr29 = getelementptr %pipe_vertex_buffer* %91, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride30 = load i32* %.stride_ptr29           ; <i32> [#uses=1]
  %.max_index_ptr31 = getelementptr %pipe_vertex_buffer* %91, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index32 = load i32* %.max_index_ptr31     ; <i32> [#uses=2]
  %.buffer_offset_ptr33 = getelementptr %pipe_vertex_buffer* %91, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset34 = load i32* %.buffer_offset_ptr33 ; <i32> [#uses=1]
  %93 = icmp ule i32 %90, %.max_index32           ; <i1> [#uses=1]
  %94 = select i1 %93, i32 %90, i32 %.max_index32 ; <i32> [#uses=1]
  %95 = mul i32 %.stride30, %94                   ; <i32> [#uses=1]
  %vbuffer35 = load i8** %92                      ; <i8*> [#uses=1]
  %96 = add i32 %95, %.buffer_offset34            ; <i32> [#uses=1]
  %97 = add i32 %96, 0                            ; <i32> [#uses=1]
  %98 = getelementptr i8* %vbuffer35, i32 %97     ; <i8*> [#uses=2]
  %src_tmp36 = getelementptr i8* %98, i32 0       ; <i8*> [#uses=1]
  %99 = bitcast i8* %src_tmp36 to float*          ; <float*> [#uses=1]
  %100 = load float* %99                          ; <float> [#uses=1]
  %101 = insertelement <4 x float> zeroinitializer, float %100, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp37 = getelementptr i8* %98, i32 4       ; <i8*> [#uses=1]
  %102 = bitcast i8* %src_tmp37 to float*         ; <float*> [#uses=1]
  %103 = load float* %102                         ; <float> [#uses=1]
  %104 = insertelement <4 x float> %101, float %103, i32 1 ; <<4 x float>> [#uses=1]
  %105 = insertelement <4 x float> %104, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %106 = insertelement <4 x float> %105, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %107 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %108 = getelementptr i8** %vbuffers, i64 0      ; <i8**> [#uses=1]
  %.stride_ptr38 = getelementptr %pipe_vertex_buffer* %107, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride39 = load i32* %.stride_ptr38           ; <i32> [#uses=1]
  %.max_index_ptr40 = getelementptr %pipe_vertex_buffer* %107, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index41 = load i32* %.max_index_ptr40     ; <i32> [#uses=2]
  %.buffer_offset_ptr42 = getelementptr %pipe_vertex_buffer* %107, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset43 = load i32* %.buffer_offset_ptr42 ; <i32> [#uses=1]
  %109 = icmp ule i32 %90, %.max_index41          ; <i1> [#uses=1]
  %110 = select i1 %109, i32 %90, i32 %.max_index41 ; <i32> [#uses=1]
  %111 = mul i32 %.stride39, %110                 ; <i32> [#uses=1]
  %vbuffer44 = load i8** %108                     ; <i8*> [#uses=1]
  %112 = add i32 %111, %.buffer_offset43          ; <i32> [#uses=1]
  %113 = add i32 %112, 8                          ; <i32> [#uses=1]
  %114 = getelementptr i8* %vbuffer44, i32 %113   ; <i8*> [#uses=2]
  %src_tmp45 = getelementptr i8* %114, i32 0      ; <i8*> [#uses=1]
  %115 = bitcast i8* %src_tmp45 to float*         ; <float*> [#uses=1]
  %116 = load float* %115                         ; <float> [#uses=1]
  %117 = insertelement <4 x float> zeroinitializer, float %116, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp46 = getelementptr i8* %114, i32 4      ; <i8*> [#uses=1]
  %118 = bitcast i8* %src_tmp46 to float*         ; <float*> [#uses=1]
  %119 = load float* %118                         ; <float> [#uses=1]
  %120 = insertelement <4 x float> %117, float %119, i32 1 ; <<4 x float>> [#uses=1]
  %121 = insertelement <4 x float> %120, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %122 = insertelement <4 x float> %121, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %123 = add i32 %21, 3                           ; <i32> [#uses=4]
  %124 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %125 = getelementptr i8** %vbuffers, i64 0      ; <i8**> [#uses=1]
  %.stride_ptr47 = getelementptr %pipe_vertex_buffer* %124, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride48 = load i32* %.stride_ptr47           ; <i32> [#uses=1]
  %.max_index_ptr49 = getelementptr %pipe_vertex_buffer* %124, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index50 = load i32* %.max_index_ptr49     ; <i32> [#uses=2]
  %.buffer_offset_ptr51 = getelementptr %pipe_vertex_buffer* %124, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset52 = load i32* %.buffer_offset_ptr51 ; <i32> [#uses=1]
  %126 = icmp ule i32 %123, %.max_index50         ; <i1> [#uses=1]
  %127 = select i1 %126, i32 %123, i32 %.max_index50 ; <i32> [#uses=1]
  %128 = mul i32 %.stride48, %127                 ; <i32> [#uses=1]
  %vbuffer53 = load i8** %125                     ; <i8*> [#uses=1]
  %129 = add i32 %128, %.buffer_offset52          ; <i32> [#uses=1]
  %130 = add i32 %129, 0                          ; <i32> [#uses=1]
  %131 = getelementptr i8* %vbuffer53, i32 %130   ; <i8*> [#uses=2]
  %src_tmp54 = getelementptr i8* %131, i32 0      ; <i8*> [#uses=1]
  %132 = bitcast i8* %src_tmp54 to float*         ; <float*> [#uses=1]
  %133 = load float* %132                         ; <float> [#uses=1]
  %134 = insertelement <4 x float> zeroinitializer, float %133, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp55 = getelementptr i8* %131, i32 4      ; <i8*> [#uses=1]
  %135 = bitcast i8* %src_tmp55 to float*         ; <float*> [#uses=1]
  %136 = load float* %135                         ; <float> [#uses=1]
  %137 = insertelement <4 x float> %134, float %136, i32 1 ; <<4 x float>> [#uses=1]
  %138 = insertelement <4 x float> %137, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %139 = insertelement <4 x float> %138, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %140 = getelementptr %pipe_vertex_buffer* %vb, i32 0 ; <%pipe_vertex_buffer*> [#uses=3]
  %141 = getelementptr i8** %vbuffers, i64 0      ; <i8**> [#uses=1]
  %.stride_ptr56 = getelementptr %pipe_vertex_buffer* %140, i32 0, i32 0 ; <i32*> [#uses=1]
  %.stride57 = load i32* %.stride_ptr56           ; <i32> [#uses=1]
  %.max_index_ptr58 = getelementptr %pipe_vertex_buffer* %140, i32 0, i32 1 ; <i32*> [#uses=1]
  %.max_index59 = load i32* %.max_index_ptr58     ; <i32> [#uses=2]
  %.buffer_offset_ptr60 = getelementptr %pipe_vertex_buffer* %140, i32 0, i32 2 ; <i32*> [#uses=1]
  %.buffer_offset61 = load i32* %.buffer_offset_ptr60 ; <i32> [#uses=1]
  %142 = icmp ule i32 %123, %.max_index59         ; <i1> [#uses=1]
  %143 = select i1 %142, i32 %123, i32 %.max_index59 ; <i32> [#uses=1]
  %144 = mul i32 %.stride57, %143                 ; <i32> [#uses=1]
  %vbuffer62 = load i8** %141                     ; <i8*> [#uses=1]
  %145 = add i32 %144, %.buffer_offset61          ; <i32> [#uses=1]
  %146 = add i32 %145, 8                          ; <i32> [#uses=1]
  %147 = getelementptr i8* %vbuffer62, i32 %146   ; <i8*> [#uses=2]
  %src_tmp63 = getelementptr i8* %147, i32 0      ; <i8*> [#uses=1]
  %148 = bitcast i8* %src_tmp63 to float*         ; <float*> [#uses=1]
  %149 = load float* %148                         ; <float> [#uses=1]
  %150 = insertelement <4 x float> zeroinitializer, float %149, i32 0 ; <<4 x float>> [#uses=1]
  %src_tmp64 = getelementptr i8* %147, i32 4      ; <i8*> [#uses=1]
  %151 = bitcast i8* %src_tmp64 to float*         ; <float*> [#uses=1]
  %152 = load float* %151                         ; <float> [#uses=1]
  %153 = insertelement <4 x float> %150, float %152, i32 1 ; <<4 x float>> [#uses=1]
  %154 = insertelement <4 x float> %153, float 0.000000e+00, i32 2 ; <<4 x float>> [#uses=1]
  %155 = insertelement <4 x float> %154, float 1.000000e+00, i32 3 ; <<4 x float>> [#uses=4]
  %156 = extractelement <4 x float> %40, i32 0    ; <float> [#uses=1]
  %157 = insertelement <4 x float> zeroinitializer, float %156, i32 0 ; <<4 x float>> [#uses=1]
  %158 = extractelement <4 x float> %73, i32 0    ; <float> [#uses=1]
  %159 = insertelement <4 x float> %157, float %158, i32 1 ; <<4 x float>> [#uses=1]
  %160 = extractelement <4 x float> %106, i32 0   ; <float> [#uses=1]
  %161 = insertelement <4 x float> %159, float %160, i32 2 ; <<4 x float>> [#uses=1]
  %162 = extractelement <4 x float> %139, i32 0   ; <float> [#uses=1]
  %163 = insertelement <4 x float> %161, float %162, i32 3 ; <<4 x float>> [#uses=4]
  %164 = extractelement <4 x float> %40, i32 1    ; <float> [#uses=1]
  %165 = insertelement <4 x float> zeroinitializer, float %164, i32 0 ; <<4 x float>> [#uses=1]
  %166 = extractelement <4 x float> %73, i32 1    ; <float> [#uses=1]
  %167 = insertelement <4 x float> %165, float %166, i32 1 ; <<4 x float>> [#uses=1]
  %168 = extractelement <4 x float> %106, i32 1   ; <float> [#uses=1]
  %169 = insertelement <4 x float> %167, float %168, i32 2 ; <<4 x float>> [#uses=1]
  %170 = extractelement <4 x float> %139, i32 1   ; <float> [#uses=1]
  %171 = insertelement <4 x float> %169, float %170, i32 3 ; <<4 x float>> [#uses=4]
  %172 = extractelement <4 x float> %40, i32 2    ; <float> [#uses=1]
  %173 = insertelement <4 x float> zeroinitializer, float %172, i32 0 ; <<4 x float>> [#uses=1]
  %174 = extractelement <4 x float> %73, i32 2    ; <float> [#uses=1]
  %175 = insertelement <4 x float> %173, float %174, i32 1 ; <<4 x float>> [#uses=1]
  %176 = extractelement <4 x float> %106, i32 2   ; <float> [#uses=1]
  %177 = insertelement <4 x float> %175, float %176, i32 2 ; <<4 x float>> [#uses=1]
  %178 = extractelement <4 x float> %139, i32 2   ; <float> [#uses=1]
  %179 = insertelement <4 x float> %177, float %178, i32 3 ; <<4 x float>> [#uses=4]
  %180 = extractelement <4 x float> %40, i32 3    ; <float> [#uses=1]
  %181 = insertelement <4 x float> zeroinitializer, float %180, i32 0 ; <<4 x float>> [#uses=1]
  %182 = extractelement <4 x float> %73, i32 3    ; <float> [#uses=1]
  %183 = insertelement <4 x float> %181, float %182, i32 1 ; <<4 x float>> [#uses=1]
  %184 = extractelement <4 x float> %106, i32 3   ; <float> [#uses=1]
  %185 = insertelement <4 x float> %183, float %184, i32 2 ; <<4 x float>> [#uses=1]
  %186 = extractelement <4 x float> %139, i32 3   ; <float> [#uses=1]
  %187 = insertelement <4 x float> %185, float %186, i32 3 ; <<4 x float>> [#uses=4]
  %188 = extractelement <4 x float> %56, i32 0    ; <float> [#uses=1]
  %189 = insertelement <4 x float> zeroinitializer, float %188, i32 0 ; <<4 x float>> [#uses=1]
  %190 = extractelement <4 x float> %89, i32 0    ; <float> [#uses=1]
  %191 = insertelement <4 x float> %189, float %190, i32 1 ; <<4 x float>> [#uses=1]
  %192 = extractelement <4 x float> %122, i32 0   ; <float> [#uses=1]
  %193 = insertelement <4 x float> %191, float %192, i32 2 ; <<4 x float>> [#uses=1]
  %194 = extractelement <4 x float> %155, i32 0   ; <float> [#uses=1]
  %195 = insertelement <4 x float> %193, float %194, i32 3 ; <<4 x float>> [#uses=1]
  %196 = extractelement <4 x float> %56, i32 1    ; <float> [#uses=1]
  %197 = insertelement <4 x float> zeroinitializer, float %196, i32 0 ; <<4 x float>> [#uses=1]
  %198 = extractelement <4 x float> %89, i32 1    ; <float> [#uses=1]
  %199 = insertelement <4 x float> %197, float %198, i32 1 ; <<4 x float>> [#uses=1]
  %200 = extractelement <4 x float> %122, i32 1   ; <float> [#uses=1]
  %201 = insertelement <4 x float> %199, float %200, i32 2 ; <<4 x float>> [#uses=1]
  %202 = extractelement <4 x float> %155, i32 1   ; <float> [#uses=1]
  %203 = insertelement <4 x float> %201, float %202, i32 3 ; <<4 x float>> [#uses=1]
  %204 = extractelement <4 x float> %56, i32 2    ; <float> [#uses=1]
  %205 = insertelement <4 x float> zeroinitializer, float %204, i32 0 ; <<4 x float>> [#uses=1]
  %206 = extractelement <4 x float> %89, i32 2    ; <float> [#uses=1]
  %207 = insertelement <4 x float> %205, float %206, i32 1 ; <<4 x float>> [#uses=1]
  %208 = extractelement <4 x float> %122, i32 2   ; <float> [#uses=1]
  %209 = insertelement <4 x float> %207, float %208, i32 2 ; <<4 x float>> [#uses=1]
  %210 = extractelement <4 x float> %155, i32 2   ; <float> [#uses=1]
  %211 = insertelement <4 x float> %209, float %210, i32 3 ; <<4 x float>> [#uses=1]
  %212 = extractelement <4 x float> %56, i32 3    ; <float> [#uses=1]
  %213 = insertelement <4 x float> zeroinitializer, float %212, i32 0 ; <<4 x float>> [#uses=1]
  %214 = extractelement <4 x float> %89, i32 3    ; <float> [#uses=1]
  %215 = insertelement <4 x float> %213, float %214, i32 1 ; <<4 x float>> [#uses=1]
  %216 = extractelement <4 x float> %122, i32 3   ; <float> [#uses=1]
  %217 = insertelement <4 x float> %215, float %216, i32 2 ; <<4 x float>> [#uses=1]
  %218 = extractelement <4 x float> %155, i32 3   ; <float> [#uses=1]
  %219 = insertelement <4 x float> %217, float %218, i32 3 ; <<4 x float>> [#uses=1]
  %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=16]
  %220 = getelementptr float* %context.vs_constants, i32 0 ; <float*> [#uses=1]
  %221 = load float* %220                         ; <float> [#uses=1]
  %222 = insertelement <4 x float> undef, float %221, i32 0 ; <<4 x float>> [#uses=1]
  %223 = shufflevector <4 x float> %222, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %224 = mul <4 x float> %163, %223               ; <<4 x float>> [#uses=1]
  %225 = getelementptr float* %context.vs_constants, i32 1 ; <float*> [#uses=1]
  %226 = load float* %225                         ; <float> [#uses=1]
  %227 = insertelement <4 x float> undef, float %226, i32 0 ; <<4 x float>> [#uses=1]
  %228 = shufflevector <4 x float> %227, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %229 = mul <4 x float> %163, %228               ; <<4 x float>> [#uses=1]
  %230 = getelementptr float* %context.vs_constants, i32 2 ; <float*> [#uses=1]
  %231 = load float* %230                         ; <float> [#uses=1]
  %232 = insertelement <4 x float> undef, float %231, i32 0 ; <<4 x float>> [#uses=1]
  %233 = shufflevector <4 x float> %232, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %234 = mul <4 x float> %163, %233               ; <<4 x float>> [#uses=1]
  %235 = getelementptr float* %context.vs_constants, i32 3 ; <float*> [#uses=1]
  %236 = load float* %235                         ; <float> [#uses=1]
  %237 = insertelement <4 x float> undef, float %236, i32 0 ; <<4 x float>> [#uses=1]
  %238 = shufflevector <4 x float> %237, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %239 = mul <4 x float> %163, %238               ; <<4 x float>> [#uses=1]
  store <4 x float> %224, <4 x float>* %11
  store <4 x float> %229, <4 x float>* %10
  store <4 x float> %234, <4 x float>* %9
  store <4 x float> %239, <4 x float>* %8
  %240 = getelementptr float* %context.vs_constants, i32 4 ; <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 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %245 = mul <4 x float> %171, %243               ; <<4 x float>> [#uses=1]
  %246 = add <4 x float> %245, %244               ; <<4 x float>> [#uses=1]
  %247 = getelementptr float* %context.vs_constants, i32 5 ; <float*> [#uses=1]
  %248 = load float* %247                         ; <float> [#uses=1]
  %249 = insertelement <4 x float> undef, float %248, i32 0 ; <<4 x float>> [#uses=1]
  %250 = shufflevector <4 x float> %249, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %251 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %252 = mul <4 x float> %171, %250               ; <<4 x float>> [#uses=1]
  %253 = add <4 x float> %252, %251               ; <<4 x float>> [#uses=1]
  %254 = getelementptr float* %context.vs_constants, i32 6 ; <float*> [#uses=1]
  %255 = load float* %254                         ; <float> [#uses=1]
  %256 = insertelement <4 x float> undef, float %255, i32 0 ; <<4 x float>> [#uses=1]
  %257 = shufflevector <4 x float> %256, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %258 = load <4 x float>* %9                     ; <<4 x float>> [#uses=1]
  %259 = mul <4 x float> %171, %257               ; <<4 x float>> [#uses=1]
  %260 = add <4 x float> %259, %258               ; <<4 x float>> [#uses=1]
  %261 = getelementptr float* %context.vs_constants, i32 7 ; <float*> [#uses=1]
  %262 = load float* %261                         ; <float> [#uses=1]
  %263 = insertelement <4 x float> undef, float %262, i32 0 ; <<4 x float>> [#uses=1]
  %264 = shufflevector <4 x float> %263, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %265 = load <4 x float>* %8                     ; <<4 x float>> [#uses=1]
  %266 = mul <4 x float> %171, %264               ; <<4 x float>> [#uses=1]
  %267 = add <4 x float> %266, %265               ; <<4 x float>> [#uses=1]
  store <4 x float> %246, <4 x float>* %11
  store <4 x float> %253, <4 x float>* %10
  store <4 x float> %260, <4 x float>* %9
  store <4 x float> %267, <4 x float>* %8
  %268 = getelementptr float* %context.vs_constants, i32 8 ; <float*> [#uses=1]
  %269 = load float* %268                         ; <float> [#uses=1]
  %270 = insertelement <4 x float> undef, float %269, i32 0 ; <<4 x float>> [#uses=1]
  %271 = shufflevector <4 x float> %270, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %272 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %273 = mul <4 x float> %179, %271               ; <<4 x float>> [#uses=1]
  %274 = add <4 x float> %273, %272               ; <<4 x float>> [#uses=1]
  %275 = getelementptr float* %context.vs_constants, i32 9 ; <float*> [#uses=1]
  %276 = load float* %275                         ; <float> [#uses=1]
  %277 = insertelement <4 x float> undef, float %276, i32 0 ; <<4 x float>> [#uses=1]
  %278 = shufflevector <4 x float> %277, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %279 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %280 = mul <4 x float> %179, %278               ; <<4 x float>> [#uses=1]
  %281 = add <4 x float> %280, %279               ; <<4 x float>> [#uses=1]
  %282 = getelementptr float* %context.vs_constants, i32 10 ; <float*> [#uses=1]
  %283 = load float* %282                         ; <float> [#uses=1]
  %284 = insertelement <4 x float> undef, float %283, i32 0 ; <<4 x float>> [#uses=1]
  %285 = shufflevector <4 x float> %284, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %286 = load <4 x float>* %9                     ; <<4 x float>> [#uses=1]
  %287 = mul <4 x float> %179, %285               ; <<4 x float>> [#uses=1]
  %288 = add <4 x float> %287, %286               ; <<4 x float>> [#uses=1]
  %289 = getelementptr float* %context.vs_constants, i32 11 ; <float*> [#uses=1]
  %290 = load float* %289                         ; <float> [#uses=1]
  %291 = insertelement <4 x float> undef, float %290, i32 0 ; <<4 x float>> [#uses=1]
  %292 = shufflevector <4 x float> %291, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %293 = load <4 x float>* %8                     ; <<4 x float>> [#uses=1]
  %294 = mul <4 x float> %179, %292               ; <<4 x float>> [#uses=1]
  %295 = add <4 x float> %294, %293               ; <<4 x float>> [#uses=1]
  store <4 x float> %274, <4 x float>* %11
  store <4 x float> %281, <4 x float>* %10
  store <4 x float> %288, <4 x float>* %9
  store <4 x float> %295, <4 x float>* %8
  %296 = getelementptr float* %context.vs_constants, i32 12 ; <float*> [#uses=1]
  %297 = load float* %296                         ; <float> [#uses=1]
  %298 = insertelement <4 x float> undef, float %297, i32 0 ; <<4 x float>> [#uses=1]
  %299 = shufflevector <4 x float> %298, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %300 = load <4 x float>* %11                    ; <<4 x float>> [#uses=1]
  %301 = mul <4 x float> %187, %299               ; <<4 x float>> [#uses=1]
  %302 = add <4 x float> %301, %300               ; <<4 x float>> [#uses=1]
  %303 = getelementptr float* %context.vs_constants, i32 13 ; <float*> [#uses=1]
  %304 = load float* %303                         ; <float> [#uses=1]
  %305 = insertelement <4 x float> undef, float %304, i32 0 ; <<4 x float>> [#uses=1]
  %306 = shufflevector <4 x float> %305, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %307 = load <4 x float>* %10                    ; <<4 x float>> [#uses=1]
  %308 = mul <4 x float> %187, %306               ; <<4 x float>> [#uses=1]
  %309 = add <4 x float> %308, %307               ; <<4 x float>> [#uses=1]
  %310 = getelementptr float* %context.vs_constants, i32 14 ; <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]
  %314 = load <4 x float>* %9                     ; <<4 x float>> [#uses=1]
  %315 = mul <4 x float> %187, %313               ; <<4 x float>> [#uses=1]
  %316 = add <4 x float> %315, %314               ; <<4 x float>> [#uses=1]
  %317 = getelementptr float* %context.vs_constants, i32 15 ; <float*> [#uses=1]
  %318 = load float* %317                         ; <float> [#uses=1]
  %319 = insertelement <4 x float> undef, float %318, i32 0 ; <<4 x float>> [#uses=1]
  %320 = shufflevector <4 x float> %319, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %321 = load <4 x float>* %8                     ; <<4 x float>> [#uses=1]
  %322 = mul <4 x float> %187, %320               ; <<4 x float>> [#uses=1]
  %323 = add <4 x float> %322, %321               ; <<4 x float>> [#uses=1]
  store <4 x float> %302, <4 x float>* %3
  store <4 x float> %309, <4 x float>* %2
  store <4 x float> %316, <4 x float>* %1
  store <4 x float> %323, <4 x float>* %0
  store <4 x float> %195, <4 x float>* %15
  store <4 x float> %203, <4 x float>* %14
  store <4 x float> %211, <4 x float>* %13
  store <4 x float> %219, <4 x float>* %12
  %324 = load <4 x float>* %3                     ; <<4 x float>> [#uses=1]
  %325 = load <4 x float>* %2                     ; <<4 x float>> [#uses=1]
  %326 = load <4 x float>* %1                     ; <<4 x float>> [#uses=1]
  %327 = load <4 x float>* %0                     ; <<4 x float>> [#uses=1]
  store <4 x float> %324, <4 x float>* %19
  store <4 x float> %325, <4 x float>* %18
  store <4 x float> %326, <4 x float>* %17
  store <4 x float> %327, <4 x float>* %16
  %328 = load <4 x float>* %3                     ; <<4 x float>> [#uses=1]
  %329 = load <4 x float>* %2                     ; <<4 x float>> [#uses=1]
  %330 = load <4 x float>* %1                     ; <<4 x float>> [#uses=1]
  %331 = load <4 x float>* %0                     ; <<4 x float>> [#uses=1]
  store <4 x float> %328, <4 x float>* %7
  store <4 x float> %329, <4 x float>* %6
  store <4 x float> %330, <4 x float>* %5
  store <4 x float> %331, <4 x float>* %4
  %output0.x = load <4 x float>* %19              ; <<4 x float>> [#uses=4]
  %output0.y = load <4 x float>* %18              ; <<4 x float>> [#uses=4]
  %output0.z = load <4 x float>* %17              ; <<4 x float>> [#uses=4]
  %output0.w = load <4 x float>* %16              ; <<4 x float>> [#uses=4]
  %332 = extractelement <4 x float> %output0.x, i32 0 ; <float> [#uses=1]
  %333 = insertelement <4 x float> zeroinitializer, float %332, i32 0 ; <<4 x float>> [#uses=1]
  %334 = extractelement <4 x float> %output0.x, i32 1 ; <float> [#uses=1]
  %335 = insertelement <4 x float> zeroinitializer, float %334, i32 0 ; <<4 x float>> [#uses=1]
  %336 = extractelement <4 x float> %output0.x, i32 2 ; <float> [#uses=1]
  %337 = insertelement <4 x float> zeroinitializer, float %336, i32 0 ; <<4 x float>> [#uses=1]
  %338 = extractelement <4 x float> %output0.x, i32 3 ; <float> [#uses=1]
  %339 = insertelement <4 x float> zeroinitializer, float %338, i32 0 ; <<4 x float>> [#uses=1]
  %340 = extractelement <4 x float> %output0.y, i32 0 ; <float> [#uses=1]
  %341 = insertelement <4 x float> %333, float %340, i32 1 ; <<4 x float>> [#uses=1]
  %342 = extractelement <4 x float> %output0.y, i32 1 ; <float> [#uses=1]
  %343 = insertelement <4 x float> %335, float %342, i32 1 ; <<4 x float>> [#uses=1]
  %344 = extractelement <4 x float> %output0.y, i32 2 ; <float> [#uses=1]
  %345 = insertelement <4 x float> %337, float %344, i32 1 ; <<4 x float>> [#uses=1]
  %346 = extractelement <4 x float> %output0.y, i32 3 ; <float> [#uses=1]
  %347 = insertelement <4 x float> %339, float %346, i32 1 ; <<4 x float>> [#uses=1]
  %348 = extractelement <4 x float> %output0.z, i32 0 ; <float> [#uses=1]
  %349 = insertelement <4 x float> %341, float %348, i32 2 ; <<4 x float>> [#uses=1]
  %350 = extractelement <4 x float> %output0.z, i32 1 ; <float> [#uses=1]
  %351 = insertelement <4 x float> %343, float %350, i32 2 ; <<4 x float>> [#uses=1]
  %352 = extractelement <4 x float> %output0.z, i32 2 ; <float> [#uses=1]
  %353 = insertelement <4 x float> %345, float %352, i32 2 ; <<4 x float>> [#uses=1]
  %354 = extractelement <4 x float> %output0.z, i32 3 ; <float> [#uses=1]
  %355 = insertelement <4 x float> %347, float %354, i32 2 ; <<4 x float>> [#uses=1]
  %356 = extractelement <4 x float> %output0.w, i32 0 ; <float> [#uses=1]
  %357 = insertelement <4 x float> %349, float %356, i32 3 ; <<4 x float>> [#uses=4]
  %358 = extractelement <4 x float> %output0.w, i32 1 ; <float> [#uses=1]
  %359 = insertelement <4 x float> %351, float %358, i32 3 ; <<4 x float>> [#uses=4]
  %360 = extractelement <4 x float> %output0.w, i32 2 ; <float> [#uses=1]
  %361 = insertelement <4 x float> %353, float %360, i32 3 ; <<4 x float>> [#uses=4]
  %362 = extractelement <4 x float> %output0.w, i32 3 ; <float> [#uses=1]
  %363 = insertelement <4 x float> %355, float %362, i32 3 ; <<4 x float>> [#uses=4]
  %364 = getelementptr %vertex_header4* %23, i32 0 ; <%vertex_header4*> [#uses=2]
  %365 = getelementptr %vertex_header4* %23, i32 1 ; <%vertex_header4*> [#uses=2]
  %366 = getelementptr %vertex_header4* %23, i32 2 ; <%vertex_header4*> [#uses=2]
  %367 = getelementptr %vertex_header4* %23, i32 3 ; <%vertex_header4*> [#uses=2]
  %.id_ptr = getelementptr %vertex_header4* %364, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr = getelementptr %vertex_header4* %364, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr
  %368 = getelementptr [4 x [4 x float]]* %.data_ptr, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %369 = extractelement <4 x float> %357, i32 0   ; <float> [#uses=1]
  %370 = extractelement <4 x float> %357, i32 1   ; <float> [#uses=1]
  %371 = extractelement <4 x float> %357, i32 2   ; <float> [#uses=1]
  %372 = extractelement <4 x float> %357, i32 3   ; <float> [#uses=1]
  %373 = getelementptr float* %368, i32 0         ; <float*> [#uses=1]
  %374 = getelementptr float* %368, i32 1         ; <float*> [#uses=1]
  %375 = getelementptr float* %368, i32 2         ; <float*> [#uses=1]
  %376 = getelementptr float* %368, i32 3         ; <float*> [#uses=1]
  store float %369, float* %373
  store float %370, float* %374
  store float %371, float* %375
  store float %372, float* %376
  %.id_ptr65 = getelementptr %vertex_header4* %365, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr66 = getelementptr %vertex_header4* %365, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr65
  %377 = getelementptr [4 x [4 x float]]* %.data_ptr66, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %378 = extractelement <4 x float> %359, i32 0   ; <float> [#uses=1]
  %379 = extractelement <4 x float> %359, i32 1   ; <float> [#uses=1]
  %380 = extractelement <4 x float> %359, i32 2   ; <float> [#uses=1]
  %381 = extractelement <4 x float> %359, i32 3   ; <float> [#uses=1]
  %382 = getelementptr float* %377, i32 0         ; <float*> [#uses=1]
  %383 = getelementptr float* %377, i32 1         ; <float*> [#uses=1]
  %384 = getelementptr float* %377, i32 2         ; <float*> [#uses=1]
  %385 = getelementptr float* %377, i32 3         ; <float*> [#uses=1]
  store float %378, float* %382
  store float %379, float* %383
  store float %380, float* %384
  store float %381, float* %385
  %.id_ptr67 = getelementptr %vertex_header4* %366, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr68 = getelementptr %vertex_header4* %366, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr67
  %386 = getelementptr [4 x [4 x float]]* %.data_ptr68, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %387 = extractelement <4 x float> %361, i32 0   ; <float> [#uses=1]
  %388 = extractelement <4 x float> %361, i32 1   ; <float> [#uses=1]
  %389 = extractelement <4 x float> %361, i32 2   ; <float> [#uses=1]
  %390 = extractelement <4 x float> %361, i32 3   ; <float> [#uses=1]
  %391 = getelementptr float* %386, i32 0         ; <float*> [#uses=1]
  %392 = getelementptr float* %386, i32 1         ; <float*> [#uses=1]
  %393 = getelementptr float* %386, i32 2         ; <float*> [#uses=1]
  %394 = getelementptr float* %386, i32 3         ; <float*> [#uses=1]
  store float %387, float* %391
  store float %388, float* %392
  store float %389, float* %393
  store float %390, float* %394
  %.id_ptr69 = getelementptr %vertex_header4* %367, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr70 = getelementptr %vertex_header4* %367, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr69
  %395 = getelementptr [4 x [4 x float]]* %.data_ptr70, i32 0, i32 0, i32 0 ; <float*> [#uses=4]
  %396 = extractelement <4 x float> %363, i32 0   ; <float> [#uses=1]
  %397 = extractelement <4 x float> %363, i32 1   ; <float> [#uses=1]
  %398 = extractelement <4 x float> %363, i32 2   ; <float> [#uses=1]
  %399 = extractelement <4 x float> %363, i32 3   ; <float> [#uses=1]
  %400 = getelementptr float* %395, i32 0         ; <float*> [#uses=1]
  %401 = getelementptr float* %395, i32 1         ; <float*> [#uses=1]
  %402 = getelementptr float* %395, i32 2         ; <float*> [#uses=1]
  %403 = getelementptr float* %395, i32 3         ; <float*> [#uses=1]
  store float %396, float* %400
  store float %397, float* %401
  store float %398, float* %402
  store float %399, float* %403
  %output1.x = load <4 x float>* %15              ; <<4 x float>> [#uses=4]
  %output1.y = load <4 x float>* %14              ; <<4 x float>> [#uses=4]
  %output1.z = load <4 x float>* %13              ; <<4 x float>> [#uses=4]
  %output1.w = load <4 x float>* %12              ; <<4 x float>> [#uses=4]
  %404 = extractelement <4 x float> %output1.x, i32 0 ; <float> [#uses=1]
  %405 = insertelement <4 x float> zeroinitializer, float %404, i32 0 ; <<4 x float>> [#uses=1]
  %406 = extractelement <4 x float> %output1.x, i32 1 ; <float> [#uses=1]
  %407 = insertelement <4 x float> zeroinitializer, float %406, i32 0 ; <<4 x float>> [#uses=1]
  %408 = extractelement <4 x float> %output1.x, i32 2 ; <float> [#uses=1]
  %409 = insertelement <4 x float> zeroinitializer, float %408, i32 0 ; <<4 x float>> [#uses=1]
  %410 = extractelement <4 x float> %output1.x, i32 3 ; <float> [#uses=1]
  %411 = insertelement <4 x float> zeroinitializer, float %410, i32 0 ; <<4 x float>> [#uses=1]
  %412 = extractelement <4 x float> %output1.y, i32 0 ; <float> [#uses=1]
  %413 = insertelement <4 x float> %405, float %412, i32 1 ; <<4 x float>> [#uses=1]
  %414 = extractelement <4 x float> %output1.y, i32 1 ; <float> [#uses=1]
  %415 = insertelement <4 x float> %407, float %414, i32 1 ; <<4 x float>> [#uses=1]
  %416 = extractelement <4 x float> %output1.y, i32 2 ; <float> [#uses=1]
  %417 = insertelement <4 x float> %409, float %416, i32 1 ; <<4 x float>> [#uses=1]
  %418 = extractelement <4 x float> %output1.y, i32 3 ; <float> [#uses=1]
  %419 = insertelement <4 x float> %411, float %418, i32 1 ; <<4 x float>> [#uses=1]
  %420 = extractelement <4 x float> %output1.z, i32 0 ; <float> [#uses=1]
  %421 = insertelement <4 x float> %413, float %420, i32 2 ; <<4 x float>> [#uses=1]
  %422 = extractelement <4 x float> %output1.z, i32 1 ; <float> [#uses=1]
  %423 = insertelement <4 x float> %415, float %422, i32 2 ; <<4 x float>> [#uses=1]
  %424 = extractelement <4 x float> %output1.z, i32 2 ; <float> [#uses=1]
  %425 = insertelement <4 x float> %417, float %424, i32 2 ; <<4 x float>> [#uses=1]
  %426 = extractelement <4 x float> %output1.z, i32 3 ; <float> [#uses=1]
  %427 = insertelement <4 x float> %419, float %426, i32 2 ; <<4 x float>> [#uses=1]
  %428 = extractelement <4 x float> %output1.w, i32 0 ; <float> [#uses=1]
  %429 = insertelement <4 x float> %421, float %428, i32 3 ; <<4 x float>> [#uses=4]
  %430 = extractelement <4 x float> %output1.w, i32 1 ; <float> [#uses=1]
  %431 = insertelement <4 x float> %423, float %430, i32 3 ; <<4 x float>> [#uses=4]
  %432 = extractelement <4 x float> %output1.w, i32 2 ; <float> [#uses=1]
  %433 = insertelement <4 x float> %425, float %432, i32 3 ; <<4 x float>> [#uses=4]
  %434 = extractelement <4 x float> %output1.w, i32 3 ; <float> [#uses=1]
  %435 = insertelement <4 x float> %427, float %434, i32 3 ; <<4 x float>> [#uses=4]
  %436 = getelementptr %vertex_header4* %23, i32 0 ; <%vertex_header4*> [#uses=2]
  %437 = getelementptr %vertex_header4* %23, i32 1 ; <%vertex_header4*> [#uses=2]
  %438 = getelementptr %vertex_header4* %23, i32 2 ; <%vertex_header4*> [#uses=2]
  %439 = getelementptr %vertex_header4* %23, i32 3 ; <%vertex_header4*> [#uses=2]
  %.id_ptr71 = getelementptr %vertex_header4* %436, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr72 = getelementptr %vertex_header4* %436, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr71
  %440 = getelementptr [4 x [4 x float]]* %.data_ptr72, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %441 = extractelement <4 x float> %429, i32 0   ; <float> [#uses=1]
  %442 = extractelement <4 x float> %429, i32 1   ; <float> [#uses=1]
  %443 = extractelement <4 x float> %429, i32 2   ; <float> [#uses=1]
  %444 = extractelement <4 x float> %429, i32 3   ; <float> [#uses=1]
  %445 = getelementptr float* %440, i32 0         ; <float*> [#uses=1]
  %446 = getelementptr float* %440, i32 1         ; <float*> [#uses=1]
  %447 = getelementptr float* %440, i32 2         ; <float*> [#uses=1]
  %448 = getelementptr float* %440, i32 3         ; <float*> [#uses=1]
  store float %441, float* %445
  store float %442, float* %446
  store float %443, float* %447
  store float %444, float* %448
  %.id_ptr73 = getelementptr %vertex_header4* %437, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr74 = getelementptr %vertex_header4* %437, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr73
  %449 = getelementptr [4 x [4 x float]]* %.data_ptr74, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %450 = extractelement <4 x float> %431, i32 0   ; <float> [#uses=1]
  %451 = extractelement <4 x float> %431, i32 1   ; <float> [#uses=1]
  %452 = extractelement <4 x float> %431, i32 2   ; <float> [#uses=1]
  %453 = extractelement <4 x float> %431, i32 3   ; <float> [#uses=1]
  %454 = getelementptr float* %449, i32 0         ; <float*> [#uses=1]
  %455 = getelementptr float* %449, i32 1         ; <float*> [#uses=1]
  %456 = getelementptr float* %449, i32 2         ; <float*> [#uses=1]
  %457 = getelementptr float* %449, i32 3         ; <float*> [#uses=1]
  store float %450, float* %454
  store float %451, float* %455
  store float %452, float* %456
  store float %453, float* %457
  %.id_ptr75 = getelementptr %vertex_header4* %438, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr76 = getelementptr %vertex_header4* %438, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr75
  %458 = getelementptr [4 x [4 x float]]* %.data_ptr76, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %459 = extractelement <4 x float> %433, i32 0   ; <float> [#uses=1]
  %460 = extractelement <4 x float> %433, i32 1   ; <float> [#uses=1]
  %461 = extractelement <4 x float> %433, i32 2   ; <float> [#uses=1]
  %462 = extractelement <4 x float> %433, i32 3   ; <float> [#uses=1]
  %463 = getelementptr float* %458, i32 0         ; <float*> [#uses=1]
  %464 = getelementptr float* %458, i32 1         ; <float*> [#uses=1]
  %465 = getelementptr float* %458, i32 2         ; <float*> [#uses=1]
  %466 = getelementptr float* %458, i32 3         ; <float*> [#uses=1]
  store float %459, float* %463
  store float %460, float* %464
  store float %461, float* %465
  store float %462, float* %466
  %.id_ptr77 = getelementptr %vertex_header4* %439, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr78 = getelementptr %vertex_header4* %439, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr77
  %467 = getelementptr [4 x [4 x float]]* %.data_ptr78, i32 0, i32 1, i32 0 ; <float*> [#uses=4]
  %468 = extractelement <4 x float> %435, i32 0   ; <float> [#uses=1]
  %469 = extractelement <4 x float> %435, i32 1   ; <float> [#uses=1]
  %470 = extractelement <4 x float> %435, i32 2   ; <float> [#uses=1]
  %471 = extractelement <4 x float> %435, i32 3   ; <float> [#uses=1]
  %472 = getelementptr float* %467, i32 0         ; <float*> [#uses=1]
  %473 = getelementptr float* %467, i32 1         ; <float*> [#uses=1]
  %474 = getelementptr float* %467, i32 2         ; <float*> [#uses=1]
  %475 = getelementptr float* %467, i32 3         ; <float*> [#uses=1]
  store float %468, float* %472
  store float %469, float* %473
  store float %470, float* %474
  store float %471, float* %475
  %output2.x = load <4 x float>* %7               ; <<4 x float>> [#uses=4]
  %output2.y = load <4 x float>* %6               ; <<4 x float>> [#uses=4]
  %output2.z = load <4 x float>* %5               ; <<4 x float>> [#uses=4]
  %output2.w = load <4 x float>* %4               ; <<4 x float>> [#uses=4]
  %476 = extractelement <4 x float> %output2.x, i32 0 ; <float> [#uses=1]
  %477 = insertelement <4 x float> zeroinitializer, float %476, i32 0 ; <<4 x float>> [#uses=1]
  %478 = extractelement <4 x float> %output2.x, i32 1 ; <float> [#uses=1]
  %479 = insertelement <4 x float> zeroinitializer, float %478, i32 0 ; <<4 x float>> [#uses=1]
  %480 = extractelement <4 x float> %output2.x, i32 2 ; <float> [#uses=1]
  %481 = insertelement <4 x float> zeroinitializer, float %480, i32 0 ; <<4 x float>> [#uses=1]
  %482 = extractelement <4 x float> %output2.x, i32 3 ; <float> [#uses=1]
  %483 = insertelement <4 x float> zeroinitializer, float %482, i32 0 ; <<4 x float>> [#uses=1]
  %484 = extractelement <4 x float> %output2.y, i32 0 ; <float> [#uses=1]
  %485 = insertelement <4 x float> %477, float %484, i32 1 ; <<4 x float>> [#uses=1]
  %486 = extractelement <4 x float> %output2.y, i32 1 ; <float> [#uses=1]
  %487 = insertelement <4 x float> %479, float %486, i32 1 ; <<4 x float>> [#uses=1]
  %488 = extractelement <4 x float> %output2.y, i32 2 ; <float> [#uses=1]
  %489 = insertelement <4 x float> %481, float %488, i32 1 ; <<4 x float>> [#uses=1]
  %490 = extractelement <4 x float> %output2.y, i32 3 ; <float> [#uses=1]
  %491 = insertelement <4 x float> %483, float %490, i32 1 ; <<4 x float>> [#uses=1]
  %492 = extractelement <4 x float> %output2.z, i32 0 ; <float> [#uses=1]
  %493 = insertelement <4 x float> %485, float %492, i32 2 ; <<4 x float>> [#uses=1]
  %494 = extractelement <4 x float> %output2.z, i32 1 ; <float> [#uses=1]
  %495 = insertelement <4 x float> %487, float %494, i32 2 ; <<4 x float>> [#uses=1]
  %496 = extractelement <4 x float> %output2.z, i32 2 ; <float> [#uses=1]
  %497 = insertelement <4 x float> %489, float %496, i32 2 ; <<4 x float>> [#uses=1]
  %498 = extractelement <4 x float> %output2.z, i32 3 ; <float> [#uses=1]
  %499 = insertelement <4 x float> %491, float %498, i32 2 ; <<4 x float>> [#uses=1]
  %500 = extractelement <4 x float> %output2.w, i32 0 ; <float> [#uses=1]
  %501 = insertelement <4 x float> %493, float %500, i32 3 ; <<4 x float>> [#uses=4]
  %502 = extractelement <4 x float> %output2.w, i32 1 ; <float> [#uses=1]
  %503 = insertelement <4 x float> %495, float %502, i32 3 ; <<4 x float>> [#uses=4]
  %504 = extractelement <4 x float> %output2.w, i32 2 ; <float> [#uses=1]
  %505 = insertelement <4 x float> %497, float %504, i32 3 ; <<4 x float>> [#uses=4]
  %506 = extractelement <4 x float> %output2.w, i32 3 ; <float> [#uses=1]
  %507 = insertelement <4 x float> %499, float %506, i32 3 ; <<4 x float>> [#uses=4]
  %508 = getelementptr %vertex_header4* %23, i32 0 ; <%vertex_header4*> [#uses=2]
  %509 = getelementptr %vertex_header4* %23, i32 1 ; <%vertex_header4*> [#uses=2]
  %510 = getelementptr %vertex_header4* %23, i32 2 ; <%vertex_header4*> [#uses=2]
  %511 = getelementptr %vertex_header4* %23, i32 3 ; <%vertex_header4*> [#uses=2]
  %.id_ptr79 = getelementptr %vertex_header4* %508, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr80 = getelementptr %vertex_header4* %508, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr79
  %512 = getelementptr [4 x [4 x float]]* %.data_ptr80, i32 0, i32 2, i32 0 ; <float*> [#uses=4]
  %513 = extractelement <4 x float> %501, i32 0   ; <float> [#uses=1]
  %514 = extractelement <4 x float> %501, i32 1   ; <float> [#uses=1]
  %515 = extractelement <4 x float> %501, i32 2   ; <float> [#uses=1]
  %516 = extractelement <4 x float> %501, i32 3   ; <float> [#uses=1]
  %517 = getelementptr float* %512, i32 0         ; <float*> [#uses=1]
  %518 = getelementptr float* %512, i32 1         ; <float*> [#uses=1]
  %519 = getelementptr float* %512, i32 2         ; <float*> [#uses=1]
  %520 = getelementptr float* %512, i32 3         ; <float*> [#uses=1]
  store float %513, float* %517
  store float %514, float* %518
  store float %515, float* %519
  store float %516, float* %520
  %.id_ptr81 = getelementptr %vertex_header4* %509, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr82 = getelementptr %vertex_header4* %509, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr81
  %521 = getelementptr [4 x [4 x float]]* %.data_ptr82, i32 0, i32 2, i32 0 ; <float*> [#uses=4]
  %522 = extractelement <4 x float> %503, i32 0   ; <float> [#uses=1]
  %523 = extractelement <4 x float> %503, i32 1   ; <float> [#uses=1]
  %524 = extractelement <4 x float> %503, i32 2   ; <float> [#uses=1]
  %525 = extractelement <4 x float> %503, i32 3   ; <float> [#uses=1]
  %526 = getelementptr float* %521, i32 0         ; <float*> [#uses=1]
  %527 = getelementptr float* %521, i32 1         ; <float*> [#uses=1]
  %528 = getelementptr float* %521, i32 2         ; <float*> [#uses=1]
  %529 = getelementptr float* %521, i32 3         ; <float*> [#uses=1]
  store float %522, float* %526
  store float %523, float* %527
  store float %524, float* %528
  store float %525, float* %529
  %.id_ptr83 = getelementptr %vertex_header4* %510, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr84 = getelementptr %vertex_header4* %510, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr83
  %530 = getelementptr [4 x [4 x float]]* %.data_ptr84, i32 0, i32 2, i32 0 ; <float*> [#uses=4]
  %531 = extractelement <4 x float> %505, i32 0   ; <float> [#uses=1]
  %532 = extractelement <4 x float> %505, i32 1   ; <float> [#uses=1]
  %533 = extractelement <4 x float> %505, i32 2   ; <float> [#uses=1]
  %534 = extractelement <4 x float> %505, i32 3   ; <float> [#uses=1]
  %535 = getelementptr float* %530, i32 0         ; <float*> [#uses=1]
  %536 = getelementptr float* %530, i32 1         ; <float*> [#uses=1]
  %537 = getelementptr float* %530, i32 2         ; <float*> [#uses=1]
  %538 = getelementptr float* %530, i32 3         ; <float*> [#uses=1]
  store float %531, float* %535
  store float %532, float* %536
  store float %533, float* %537
  store float %534, float* %538
  %.id_ptr85 = getelementptr %vertex_header4* %511, i32 0, i32 0 ; <i32*> [#uses=1]
  %.data_ptr86 = getelementptr %vertex_header4* %511, i32 0, i32 2 ; <[4 x [4 x float]]*> [#uses=1]
  store i32 65535, i32* %.id_ptr85
  %539 = getelementptr [4 x [4 x float]]* %.data_ptr86, i32 0, i32 2, i32 0 ; <float*> [#uses=4]
  %540 = extractelement <4 x float> %507, i32 0   ; <float> [#uses=1]
  %541 = extractelement <4 x float> %507, i32 1   ; <float> [#uses=1]
  %542 = extractelement <4 x float> %507, i32 2   ; <float> [#uses=1]
  %543 = extractelement <4 x float> %507, i32 3   ; <float> [#uses=1]
  %544 = getelementptr float* %539, i32 0         ; <float*> [#uses=1]
  %545 = getelementptr float* %539, i32 1         ; <float*> [#uses=1]
  %546 = getelementptr float* %539, i32 2         ; <float*> [#uses=1]
  %547 = getelementptr float* %539, i32 3         ; <float*> [#uses=1]
  store float %540, float* %544
  store float %541, float* %545
  store float %542, float* %546
  store float %543, float* %547
  %548 = add i32 %21, 4                           ; <i32> [#uses=2]
  %549 = icmp uge i32 %548, %20                   ; <i1> [#uses=1]
  br i1 %549, label %550, label %loop

; <label>:550                                     ; 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: 0x1010550: f32 = add 0xfc40b0, 0x100f040 [ID=524]
  0xfc40b0: f32 = mul 0xfc49c0, 0xfe9300 [ID=353]
    0xfc49c0: f32 = extract_vector_elt 0xfe6be0, 0xfbf060 [ID=282]
      0xfe6be0: v4f32 = vector_shuffle 0x1004050, 0xfde480<0,0,0,0> [ORD=270] [ID=243]
        0x1004050: v4f32 = scalar_to_vector 0xfe69e0 [ID=223]
          0xfe69e0: f32,ch = load 0xd3c9e8, 0xfe68e0, 0xfb3d30<LD4[%222]> [ORD=268] [ID=152]
            0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
            0xfe68e0: i64 = add 0xfde280, 0xfe67e0 [ORD=267] [ID=130]
              0xfde280: i64,ch = load 0xd3c9e8, 0xfde180, 0xfb3d30<LD8[%context.vs_constants_ptr136]> [ORD=198] [ID=114]
                0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                0xfde180: i64,ch = CopyFromReg 0xd3c9e8, 0xfde080 [ORD=197] [ID=100]
                  0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                  0xfde080: i64 = Register %reg1024 [ORD=197] [ID=19]
                0xfb3d30: i64 = undef [ORD=35] [ID=15]
              0xfe67e0: i64 = Constant<48> [ORD=267] [ID=31]
            0xfb3d30: i64 = undef [ORD=35] [ID=15]
        0xfde480: v4f32 = undef [ORD=201] [ID=20]
      0xfbf060: i32 = Constant<0> [ID=12]
    0xfe9300: f32,ch = load 0xd3c9e8, 0xfe73e0, 0xfb3d30<LD4[ConstantPool]> [ID=115]
      0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
      0xfe73e0: i64 = X86ISD::Wrapper 0xfe8c00 [ID=103]
        0xfe8c00: i64 = TargetConstantPool<float 1.000000e+00> 0 [ID=86]
      0xfb3d30: i64 = undef [ORD=35] [ID=15]
  0x100f040: f32 = add 0xfea510, 0x1004b60 [ID=508]
    0xfea510: f32 = mul 0xfe9d10, 0xfb3930 [ID=369]
      0xfe9d10: f32 = extract_vector_elt 0xfc5dd0, 0xfbf060 [ID=298]
        0xfc5dd0: v4f32 = vector_shuffle 0x1001310, 0xfde480<0,0,0,0> [ORD=246] [ID=247]
          0x1001310: v4f32 = scalar_to_vector 0xfc5bd0 [ID=227]
            0xfc5bd0: f32,ch = load 0xd3c9e8, 0xfc5ad0, 0xfb3d30<LD4[%198]> [ORD=244] [ID=156]
              0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
              0xfc5ad0: i64 = add 0xfde280, 0xfc59d0 [ORD=243] [ID=134]
                0xfde280: i64,ch = load 0xd3c9e8, 0xfde180, 0xfb3d30<LD8[%context.vs_constants_ptr136]> [ORD=198] [ID=114]
                  0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                  0xfde180: i64,ch = CopyFromReg 0xd3c9e8, 0xfde080 [ORD=197] [ID=100]
                    0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                    0xfde080: i64 = Register %reg1024 [ORD=197] [ID=19]
                  0xfb3d30: i64 = undef [ORD=35] [ID=15]
                0xfc59d0: i64 = Constant<32> [ORD=243] [ID=27]
              0xfb3d30: i64 = undef [ORD=35] [ID=15]
          0xfde480: v4f32 = undef [ORD=201] [ID=20]
        0xfbf060: i32 = Constant<0> [ID=12]
      0xfb3930: f32 = ConstantFP<0.000000e+00> [ORD=36] [ID=16]
    0x1004b60: f32 = add 0x1001510, 0xffcef0 [ID=491]
      0x1001510: f32 = mul 0xfc0c70, 0x1001410 [ID=474]
        0xfc0c70: f32,ch = load 0xd3c9e8, 0xfb2d20, 0xfb3d30<LD4[%9]> [ORD=39] [ID=450]
          0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
          0xfb2d20: i64 = add 0xfb4130, 0xfb3020 [ORD=37] [ID=401]
            0xfb4130: i64 = add 0xfb4740, 0xfb4430 [ORD=32] [ID=337]
              0xfb4740: i64,ch = CopyFromReg 0xd3c9e8, 0xfb4840 [ORD=32] [ID=99]
                0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                0xfb4840: i64 = Register %reg1030 [ORD=32] [ID=13]
              0xfb4430: i64 = sign_extend 0xfbf260 [ORD=32] [ID=266]
                0xfbf260: i32 = add 0xfbf860, 0xfbf460 [ORD=30] [ID=239]
                  0xfbf860: i32 = mul 0xfbfd60, 0xfc0870 [ORD=29] [ID=168]
                    0xfbfd60: i32,ch = CopyFromReg 0xd3c9e8, 0xfb4940 [ORD=29] [ID=97]
                      0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                      0xfb4940: i32 = Register %reg1028 [ORD=29] [ID=10]
                    0xfc0870: i32 = xor 0xfe20a0, 0xfbfa60 [ID=146]
                      0xfe20a0: i32,flag = X86ISD::CMOV 0xfb2a20, 0xfb4530, 0xfb3420, 0xfc68e0 [ID=122]
                        0xfb2a20: i32,ch = CopyFromReg 0xd3c9e8, 0xfb3730 [ORD=16] [ID=91]
                          0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                          0xfb3730: i32 = Register %reg1032 [ORD=16] [ID=2]
                        0xfb4530: i32,ch = CopyFromReg 0xd3c9e8, 0xfb4230 [ORD=25] [ID=94]
                          0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                          0xfb4230: i32 = Register %reg1036 [ORD=25] [ID=6]
                        0xfb3420: i8 = Constant<0> [ID=87]
                        0xfc68e0: i32 = X86ISD::CMP 0xfb4530, 0xfb2a20 [ID=110]
                          0xfb4530: i32,ch = CopyFromReg 0xd3c9e8, 0xfb4230 [ORD=25] [ID=94]
                            0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                            0xfb4230: i32 = Register %reg1036 [ORD=25] [ID=6]
                          0xfb2a20: i32,ch = CopyFromReg 0xd3c9e8, 0xfb3730 [ORD=16] [ID=91]
                            0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                            0xfb3730: i32 = Register %reg1032 [ORD=16] [ID=2]
                      0xfbfa60: i32 = Constant<-1> [ORD=18] [ID=3]
                  0xfbf460: i32,ch = CopyFromReg 0xd3c9e8, 0xfbf660 [ORD=30] [ID=98]
                    0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                    0xfbf660: i32 = Register %reg1029 [ORD=30] [ID=11]
            0xfb3020: i64 = Constant<4> [ORD=37] [ID=17]
          0xfb3d30: i64 = undef [ORD=35] [ID=15]
        0x1001410: f32 = extract_vector_elt 0xfe1fa0, 0xfbf060 [ID=314]
          0xfe1fa0: v4f32 = vector_shuffle 0xfc5fd0, 0xfde480<0,0,0,0> [ORD=222] [ID=251]
            0xfc5fd0: v4f32 = scalar_to_vector 0xfe1da0 [ID=231]
              0xfe1da0: f32,ch = load 0xd3c9e8, 0xfe1ca0, 0xfb3d30<LD4[%174]> [ORD=220] [ID=160]
                0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                0xfe1ca0: i64 = add 0xfde280, 0xfe1ba0 [ORD=219] [ID=138]
                  0xfde280: i64,ch = load 0xd3c9e8, 0xfde180, 0xfb3d30<LD8[%context.vs_constants_ptr136]> [ORD=198] [ID=114]
                    0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                    0xfde180: i64,ch = CopyFromReg 0xd3c9e8, 0xfde080 [ORD=197] [ID=100]
                      0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                      0xfde080: i64 = Register %reg1024 [ORD=197] [ID=19]
                    0xfb3d30: i64 = undef [ORD=35] [ID=15]
                  0xfe1ba0: i64 = Constant<16> [ORD=219] [ID=23]
                0xfb3d30: i64 = undef [ORD=35] [ID=15]
            0xfde480: v4f32 = undef [ORD=201] [ID=20]
          0xfbf060: i32 = Constant<0> [ID=12]
      0xffcef0: f32 = mul 0xfb3b30, 0xffcdf0 [ID=454]
        0xfb3b30: f32,ch = load 0xd3c9e8, 0xfb4130, 0xfb3d30<LD4[%6]> [ORD=35] [ID=402]
          0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
          0xfb4130: i64 = add 0xfb4740, 0xfb4430 [ORD=32] [ID=337]
            0xfb4740: i64,ch = CopyFromReg 0xd3c9e8, 0xfb4840 [ORD=32] [ID=99]
              0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
              0xfb4840: i64 = Register %reg1030 [ORD=32] [ID=13]
            0xfb4430: i64 = sign_extend 0xfbf260 [ORD=32] [ID=266]
              0xfbf260: i32 = add 0xfbf860, 0xfbf460 [ORD=30] [ID=239]
                0xfbf860: i32 = mul 0xfbfd60, 0xfc0870 [ORD=29] [ID=168]
                  0xfbfd60: i32,ch = CopyFromReg 0xd3c9e8, 0xfb4940 [ORD=29] [ID=97]
                    0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                    0xfb4940: i32 = Register %reg1028 [ORD=29] [ID=10]
                  0xfc0870: i32 = xor 0xfe20a0, 0xfbfa60 [ID=146]
                    0xfe20a0: i32,flag = X86ISD::CMOV 0xfb2a20, 0xfb4530, 0xfb3420, 0xfc68e0 [ID=122]
                      0xfb2a20: i32,ch = CopyFromReg 0xd3c9e8, 0xfb3730 [ORD=16] [ID=91]
                        0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                        0xfb3730: i32 = Register %reg1032 [ORD=16] [ID=2]
                      0xfb4530: i32,ch = CopyFromReg 0xd3c9e8, 0xfb4230 [ORD=25] [ID=94]
                        0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                        0xfb4230: i32 = Register %reg1036 [ORD=25] [ID=6]
                      0xfb3420: i8 = Constant<0> [ID=87]
                      0xfc68e0: i32 = X86ISD::CMP 0xfb4530, 0xfb2a20 [ID=110]
                        0xfb4530: i32,ch = CopyFromReg 0xd3c9e8, 0xfb4230 [ORD=25] [ID=94]
                          0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                          0xfb4230: i32 = Register %reg1036 [ORD=25] [ID=6]
                        0xfb2a20: i32,ch = CopyFromReg 0xd3c9e8, 0xfb3730 [ORD=16] [ID=91]
                          0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                          0xfb3730: i32 = Register %reg1032 [ORD=16] [ID=2]
                    0xfbfa60: i32 = Constant<-1> [ORD=18] [ID=3]
                0xfbf460: i32,ch = CopyFromReg 0xd3c9e8, 0xfbf660 [ORD=30] [ID=98]
                  0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                  0xfbf660: i32 = Register %reg1029 [ORD=30] [ID=11]
          0xfb3d30: i64 = undef [ORD=35] [ID=15]
        0xffcdf0: f32 = extract_vector_elt 0xfc67e0, 0xfbf060 [ID=258]
          0xfc67e0: v4f32 = vector_shuffle 0xfe21a0, 0xfde480<0,0,0,0> [ORD=202] [ID=235]
            0xfe21a0: v4f32 = scalar_to_vector 0xfde380 [ID=164]
              0xfde380: f32,ch = load 0xd3c9e8, 0xfde280, 0xfb3d30<LD4[%154]> [ORD=200] [ID=142]
                0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                0xfde280: i64,ch = load 0xd3c9e8, 0xfde180, 0xfb3d30<LD8[%context.vs_constants_ptr136]> [ORD=198] [ID=114]
                  0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                  0xfde180: i64,ch = CopyFromReg 0xd3c9e8, 0xfde080 [ORD=197] [ID=100]
                    0xd3c9e8: ch = EntryToken [ORD=16] [ID=0]
                    0xfde080: i64 = Register %reg1024 [ORD=197] [ID=19]
                  0xfb3d30: i64 = undef [ORD=35] [ID=15]
                0xfb3d30: i64 = undef [ORD=35] [ID=15]
            0xfde480: v4f32 = undef [ORD=201] [ID=20]
          0xfbf060: i32 = Constant<0> [ID=12]