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
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_flags_option: GALLIVM_DEBUG = 0x0
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: GALLIVM_DEBUG = 0x0
Integer arithmetic operators only work with integral types!
  %pos.x.a = add <4 x float> %24, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00)> ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %pos.y.a = add <4 x float> %29, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00), float add (float 1.000000e+00, float 1.000000e+00)> ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %pos.z.dadxy = add float %pos.z.dadx, %pos.z.dady ; <float> [#uses=1]
Instruction does not dominate all uses!
  %pos.z.dadxy = add float %pos.z.dadx, %pos.z.dady ; <float> [#uses=1]
  %pos.z.dadq = insertelement <4 x float> %34, float %pos.z.dadxy, i32 3 ; <<4 x float>> [#uses=6]
Integer arithmetic operators only work with integral types!
  %35 = add <4 x float> %pos.z.dadq, %pos.z.dadq  ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %36 = mul float %18, %pos.z.dadx                ; <float> [#uses=1]
Integer arithmetic operators only work with integral types!
  %37 = add float %pos.z.a0, %36                  ; <float> [#uses=1]
Integer arithmetic operators only work with integral types!
  %38 = mul float %19, %pos.z.dady                ; <float> [#uses=1]
Integer arithmetic operators only work with integral types!
  %39 = add float %37, %38                        ; <float> [#uses=4]
Instruction does not dominate all uses!
  %39 = add float %37, %38                        ; <float> [#uses=4]
  %40 = insertelement <4 x float> undef, float %39, i32 0 ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %40 = insertelement <4 x float> undef, float %39, i32 0 ; <<4 x float>> [#uses=1]
  %41 = insertelement <4 x float> %40, float %39, i32 1 ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %41 = insertelement <4 x float> %40, float %39, i32 1 ; <<4 x float>> [#uses=1]
  %42 = insertelement <4 x float> %41, float %39, i32 2 ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %42 = insertelement <4 x float> %41, float %39, i32 2 ; <<4 x float>> [#uses=1]
  %43 = insertelement <4 x float> %42, float %39, i32 3 ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %pos.z.a = add <4 x float> %43, %35             ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %pos.w.dadxy = add float %pos.w.dadx, %pos.w.dady ; <float> [#uses=1]
Instruction does not dominate all uses!
  %pos.w.dadxy = add float %pos.w.dadx, %pos.w.dady ; <float> [#uses=1]
  %pos.w.dadq = insertelement <4 x float> %48, float %pos.w.dadxy, i32 3 ; <<4 x float>> [#uses=6]
Integer arithmetic operators only work with integral types!
  %49 = add <4 x float> %pos.w.dadq, %pos.w.dadq  ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %50 = mul float %18, %pos.w.dadx                ; <float> [#uses=1]
Integer arithmetic operators only work with integral types!
  %51 = add float %pos.w.a0, %50                  ; <float> [#uses=1]
Integer arithmetic operators only work with integral types!
  %52 = mul float %19, %pos.w.dady                ; <float> [#uses=1]
Integer arithmetic operators only work with integral types!
  %53 = add float %51, %52                        ; <float> [#uses=4]
Instruction does not dominate all uses!
  %53 = add float %51, %52                        ; <float> [#uses=4]
  %54 = insertelement <4 x float> undef, float %53, i32 0 ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %54 = insertelement <4 x float> undef, float %53, i32 0 ; <<4 x float>> [#uses=1]
  %55 = insertelement <4 x float> %54, float %53, i32 1 ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %55 = insertelement <4 x float> %54, float %53, i32 1 ; <<4 x float>> [#uses=1]
  %56 = insertelement <4 x float> %55, float %53, i32 2 ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %56 = insertelement <4 x float> %55, float %53, i32 2 ; <<4 x float>> [#uses=1]
  %57 = insertelement <4 x float> %56, float %53, i32 3 ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %pos.w.a = add <4 x float> %57, %49             ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %input0.x.a = add <4 x float> %62, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %input0.y.a = add <4 x float> %67, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %input0.z.a = add <4 x float> %72, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %input0.w.a = add <4 x float> %77, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
Instruction does not dominate all uses!
  %pos.x.a = add <4 x float> %24, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00)> ; <<4 x float>> [#uses=4]
  %78 = shufflevector <4 x float> %pos.x.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %pos.x = add <4 x float> %78, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Instruction does not dominate all uses!
  %pos.y.a = add <4 x float> %29, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00), float add (float 1.000000e+00, float 1.000000e+00)> ; <<4 x float>> [#uses=4]
  %79 = shufflevector <4 x float> %pos.y.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %pos.y = add <4 x float> %79, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Instruction does not dominate all uses!
  %pos.z.a = add <4 x float> %43, %35             ; <<4 x float>> [#uses=4]
  %80 = shufflevector <4 x float> %pos.z.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %pos.z = add <4 x float> %80, %pos.z.dadq       ; <<4 x float>> [#uses=4]
Instruction does not dominate all uses!
  %pos.w.a = add <4 x float> %57, %49             ; <<4 x float>> [#uses=4]
  %81 = shufflevector <4 x float> %pos.w.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %pos.w = add <4 x float> %81, %pos.w.dadq       ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %84 = mul <4 x float> %pos.z, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %85 = add <4 x float> %84, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %85 = add <4 x float> %84, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %86 = bitcast <4 x float> %85 to <4 x i32>      ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %86 = bitcast <4 x float> %85 to <4 x i32>      ; <<4 x i32>> [#uses=1]
  %87 = shl <4 x i32> %86, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %87 = shl <4 x i32> %86, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %88 = lshr <4 x i32> %87, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %88 = lshr <4 x i32> %87, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %89 = and <4 x i32> %88, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %89 = and <4 x i32> %88, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %90 = icmp ult <4 x i32> %89, %zsbuf.z          ; <<4 x i1>> [#uses=1]
Instruction does not dominate all uses!
  %90 = icmp ult <4 x i32> %89, %zsbuf.z          ; <<4 x i1>> [#uses=1]
  %91 = sext <4 x i1> %90 to <4 x i32>            ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %91 = sext <4 x i1> %90 to <4 x i32>            ; <<4 x i32>> [#uses=2]
  %92 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %91 ; <<4 x i32>> [#uses=4]
Instruction does not dominate all uses!
  %92 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %91 ; <<4 x i32>> [#uses=4]
  %93 = bitcast <4 x i32> %92 to i128             ; <i128> [#uses=1]
Instruction does not dominate all uses!
  %93 = bitcast <4 x i32> %92 to i128             ; <i128> [#uses=1]
  %94 = icmp eq i128 %93, 0                       ; <i1> [#uses=1]
Instruction does not dominate all uses!
  %94 = icmp eq i128 %93, 0                       ; <i1> [#uses=1]
  br i1 %94, label %106, label %95
Integer arithmetic operators only work with integral types!
  %pos.x1 = add <4 x float> %113, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %pos.y2 = add <4 x float> %114, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %pos.z3 = add <4 x float> %115, %pos.z.dadq     ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %pos.w4 = add <4 x float> %116, %pos.w.dadq     ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %119 = mul <4 x float> %pos.z3, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %120 = add <4 x float> %119, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %120 = add <4 x float> %119, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %121 = bitcast <4 x float> %120 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %121 = bitcast <4 x float> %120 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %122 = shl <4 x i32> %121, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %122 = shl <4 x i32> %121, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %123 = lshr <4 x i32> %122, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %123 = lshr <4 x i32> %122, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %124 = and <4 x i32> %123, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %124 = and <4 x i32> %123, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %125 = icmp ult <4 x i32> %124, %zsbuf.z12      ; <<4 x i1>> [#uses=1]
Instruction does not dominate all uses!
  %125 = icmp ult <4 x i32> %124, %zsbuf.z12      ; <<4 x i1>> [#uses=1]
  %126 = sext <4 x i1> %125 to <4 x i32>          ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %126 = sext <4 x i1> %125 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %127 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %126 ; <<4 x i32>> [#uses=4]
Instruction does not dominate all uses!
  %127 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %126 ; <<4 x i32>> [#uses=4]
  %128 = bitcast <4 x i32> %127 to i128           ; <i128> [#uses=1]
Instruction does not dominate all uses!
  %128 = bitcast <4 x i32> %127 to i128           ; <i128> [#uses=1]
  %129 = icmp eq i128 %128, 0                     ; <i1> [#uses=1]
Instruction does not dominate all uses!
  %129 = icmp eq i128 %128, 0                     ; <i1> [#uses=1]
  br i1 %129, label %141, label %130
Integer arithmetic operators only work with integral types!
  %pos.x15 = add <4 x float> %148, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %pos.y16 = add <4 x float> %149, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %pos.z17 = add <4 x float> %150, %pos.z.dadq    ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %pos.w18 = add <4 x float> %151, %pos.w.dadq    ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %154 = mul <4 x float> %pos.z17, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %155 = add <4 x float> %154, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %155 = add <4 x float> %154, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %156 = bitcast <4 x float> %155 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %156 = bitcast <4 x float> %155 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %157 = shl <4 x i32> %156, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %157 = shl <4 x i32> %156, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %158 = lshr <4 x i32> %157, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %158 = lshr <4 x i32> %157, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %159 = and <4 x i32> %158, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %159 = and <4 x i32> %158, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %160 = icmp ult <4 x i32> %159, %zsbuf.z26      ; <<4 x i1>> [#uses=1]
Instruction does not dominate all uses!
  %160 = icmp ult <4 x i32> %159, %zsbuf.z26      ; <<4 x i1>> [#uses=1]
  %161 = sext <4 x i1> %160 to <4 x i32>          ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %161 = sext <4 x i1> %160 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %162 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %161 ; <<4 x i32>> [#uses=4]
Instruction does not dominate all uses!
  %162 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %161 ; <<4 x i32>> [#uses=4]
  %163 = bitcast <4 x i32> %162 to i128           ; <i128> [#uses=1]
Instruction does not dominate all uses!
  %163 = bitcast <4 x i32> %162 to i128           ; <i128> [#uses=1]
  %164 = icmp eq i128 %163, 0                     ; <i1> [#uses=1]
Instruction does not dominate all uses!
  %164 = icmp eq i128 %163, 0                     ; <i1> [#uses=1]
  br i1 %164, label %176, label %165
Integer arithmetic operators only work with integral types!
  %pos.x29 = add <4 x float> %183, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %pos.y30 = add <4 x float> %184, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %pos.z31 = add <4 x float> %185, %pos.z.dadq    ; <<4 x float>> [#uses=4]
Integer arithmetic operators only work with integral types!
  %pos.w32 = add <4 x float> %186, %pos.w.dadq    ; <<4 x float>> [#uses=0]
Integer arithmetic operators only work with integral types!
  %189 = mul <4 x float> %pos.z31, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %190 = add <4 x float> %189, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %190 = add <4 x float> %189, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %191 = bitcast <4 x float> %190 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %191 = bitcast <4 x float> %190 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %192 = shl <4 x i32> %191, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %192 = shl <4 x i32> %191, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %193 = lshr <4 x i32> %192, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %193 = lshr <4 x i32> %192, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %194 = and <4 x i32> %193, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %194 = and <4 x i32> %193, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %195 = icmp ult <4 x i32> %194, %zsbuf.z40      ; <<4 x i1>> [#uses=1]
Instruction does not dominate all uses!
  %195 = icmp ult <4 x i32> %194, %zsbuf.z40      ; <<4 x i1>> [#uses=1]
  %196 = sext <4 x i1> %195 to <4 x i32>          ; <<4 x i32>> [#uses=2]
Instruction does not dominate all uses!
  %196 = sext <4 x i1> %195 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %197 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %196 ; <<4 x i32>> [#uses=4]
Instruction does not dominate all uses!
  %197 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %196 ; <<4 x i32>> [#uses=4]
  %198 = bitcast <4 x i32> %197 to i128           ; <i128> [#uses=1]
Instruction does not dominate all uses!
  %198 = bitcast <4 x i32> %197 to i128           ; <i128> [#uses=1]
  %199 = icmp eq i128 %198, 0                     ; <i1> [#uses=1]
Instruction does not dominate all uses!
  %199 = icmp eq i128 %198, 0                     ; <i1> [#uses=1]
  br i1 %199, label %211, label %200
Integer arithmetic operators only work with integral types!
  %226 = mul <4 x float> %222, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %227 = add <4 x float> %226, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %227 = add <4 x float> %226, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %228 = bitcast <4 x float> %227 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %228 = bitcast <4 x float> %227 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %229 = and <4 x i32> %228, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %230 = mul <4 x float> %223, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %231 = add <4 x float> %230, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %231 = add <4 x float> %230, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %232 = bitcast <4 x float> %231 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %232 = bitcast <4 x float> %231 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %233 = and <4 x i32> %232, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %234 = mul <4 x float> %224, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %235 = add <4 x float> %234, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %235 = add <4 x float> %234, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %236 = bitcast <4 x float> %235 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %236 = bitcast <4 x float> %235 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %237 = and <4 x i32> %236, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %238 = mul <4 x float> %225, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %239 = add <4 x float> %238, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %239 = add <4 x float> %238, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %240 = bitcast <4 x float> %239 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %240 = bitcast <4 x float> %239 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %241 = and <4 x i32> %240, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %229 = and <4 x i32> %228, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %242 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %229, <4 x i32> %233) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %237 = and <4 x i32> %236, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %243 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %237, <4 x i32> %241) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %242 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %229, <4 x i32> %233) ; <<8 x i16>> [#uses=1]
  %res.r = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %242, <8 x i16> %243) ; <<16 x i8>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %252 = mul <4 x float> %248, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %253 = add <4 x float> %252, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %253 = add <4 x float> %252, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %254 = bitcast <4 x float> %253 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %254 = bitcast <4 x float> %253 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %255 = and <4 x i32> %254, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %256 = mul <4 x float> %249, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %257 = add <4 x float> %256, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %257 = add <4 x float> %256, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %258 = bitcast <4 x float> %257 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %258 = bitcast <4 x float> %257 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %259 = and <4 x i32> %258, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %260 = mul <4 x float> %250, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %261 = add <4 x float> %260, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %261 = add <4 x float> %260, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %262 = bitcast <4 x float> %261 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %262 = bitcast <4 x float> %261 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %263 = and <4 x i32> %262, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %264 = mul <4 x float> %251, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %265 = add <4 x float> %264, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %265 = add <4 x float> %264, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %266 = bitcast <4 x float> %265 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %266 = bitcast <4 x float> %265 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %267 = and <4 x i32> %266, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %255 = and <4 x i32> %254, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %268 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %255, <4 x i32> %259) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %263 = and <4 x i32> %262, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %269 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %263, <4 x i32> %267) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %268 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %255, <4 x i32> %259) ; <<8 x i16>> [#uses=1]
  %res.g = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %268, <8 x i16> %269) ; <<16 x i8>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %278 = mul <4 x float> %274, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %279 = add <4 x float> %278, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %279 = add <4 x float> %278, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %280 = bitcast <4 x float> %279 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %280 = bitcast <4 x float> %279 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %281 = and <4 x i32> %280, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %282 = mul <4 x float> %275, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %283 = add <4 x float> %282, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %283 = add <4 x float> %282, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %284 = bitcast <4 x float> %283 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %284 = bitcast <4 x float> %283 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %285 = and <4 x i32> %284, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %286 = mul <4 x float> %276, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %287 = add <4 x float> %286, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %287 = add <4 x float> %286, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %288 = bitcast <4 x float> %287 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %288 = bitcast <4 x float> %287 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %289 = and <4 x i32> %288, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %290 = mul <4 x float> %277, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %291 = add <4 x float> %290, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %291 = add <4 x float> %290, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %292 = bitcast <4 x float> %291 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %292 = bitcast <4 x float> %291 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %293 = and <4 x i32> %292, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %281 = and <4 x i32> %280, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %294 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %281, <4 x i32> %285) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %289 = and <4 x i32> %288, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %295 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %289, <4 x i32> %293) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %294 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %281, <4 x i32> %285) ; <<8 x i16>> [#uses=1]
  %res.b = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %294, <8 x i16> %295) ; <<16 x i8>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %304 = mul <4 x float> %300, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %305 = add <4 x float> %304, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %305 = add <4 x float> %304, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %306 = bitcast <4 x float> %305 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %306 = bitcast <4 x float> %305 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %307 = and <4 x i32> %306, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %308 = mul <4 x float> %301, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %309 = add <4 x float> %308, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %309 = add <4 x float> %308, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %310 = bitcast <4 x float> %309 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %310 = bitcast <4 x float> %309 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %311 = and <4 x i32> %310, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %312 = mul <4 x float> %302, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %313 = add <4 x float> %312, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %313 = add <4 x float> %312, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %314 = bitcast <4 x float> %313 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %314 = bitcast <4 x float> %313 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %315 = and <4 x i32> %314, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %316 = mul <4 x float> %303, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
Integer arithmetic operators only work with integral types!
  %317 = add <4 x float> %316, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
Instruction does not dominate all uses!
  %317 = add <4 x float> %316, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %318 = bitcast <4 x float> %317 to <4 x i32>    ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %318 = bitcast <4 x float> %317 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %319 = and <4 x i32> %318, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
Instruction does not dominate all uses!
  %307 = and <4 x i32> %306, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %320 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %307, <4 x i32> %311) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %315 = and <4 x i32> %314, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %321 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %315, <4 x i32> %319) ; <<8 x i16>> [#uses=1]
Instruction does not dominate all uses!
  %320 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %307, <4 x i32> %311) ; <<8 x i16>> [#uses=1]
  %color0.a = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %320, <8 x i16> %321) ; <<16 x i8>> [#uses=0]
Broken module found, verification continues.

define void @fs4_variant0_whole(%context* noalias %context, i32 %x, i32 %y, float, float* noalias %a0, float* noalias %dadx, float* noalias %dady, <16 x i8>** noalias %color_ptr_ptr, <4 x i32>* noalias %depth, i32 %mask_input, i32* noalias) {
entry:
  %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=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=2]
  %9 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %10 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %11 = alloca <4 x float>                        ; <<4 x float>*> [#uses=2]
  %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 = sitofp i32 %x to float                    ; <float> [#uses=6]
  %19 = sitofp i32 %y to float                    ; <float> [#uses=6]
  %20 = getelementptr float* %a0, i32 0           ; <float*> [#uses=1]
  %pos.x.a0 = load float* %20                     ; <float> [#uses=0]
  %21 = insertelement <4 x float> undef, float %18, i32 0 ; <<4 x float>> [#uses=1]
  %22 = insertelement <4 x float> %21, float %18, i32 1 ; <<4 x float>> [#uses=1]
  %23 = insertelement <4 x float> %22, float %18, i32 2 ; <<4 x float>> [#uses=1]
  %24 = insertelement <4 x float> %23, float %18, i32 3 ; <<4 x float>> [#uses=1]
  %pos.x.a = add <4 x float> %24, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00)> ; <<4 x float>> [#uses=4]
  %25 = getelementptr float* %a0, i32 1           ; <float*> [#uses=1]
  %pos.y.a0 = load float* %25                     ; <float> [#uses=0]
  %26 = insertelement <4 x float> undef, float %19, i32 0 ; <<4 x float>> [#uses=1]
  %27 = insertelement <4 x float> %26, float %19, i32 1 ; <<4 x float>> [#uses=1]
  %28 = insertelement <4 x float> %27, float %19, i32 2 ; <<4 x float>> [#uses=1]
  %29 = insertelement <4 x float> %28, float %19, i32 3 ; <<4 x float>> [#uses=1]
  %pos.y.a = add <4 x float> %29, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 1.000000e+00, float 1.000000e+00), float add (float 1.000000e+00, float 1.000000e+00)> ; <<4 x float>> [#uses=4]
  %30 = getelementptr float* %dadx, i32 2         ; <float*> [#uses=1]
  %pos.z.dadx = load float* %30                   ; <float> [#uses=3]
  %31 = getelementptr float* %dady, i32 2         ; <float*> [#uses=1]
  %pos.z.dady = load float* %31                   ; <float> [#uses=3]
  %pos.z.dadxy = add float %pos.z.dadx, %pos.z.dady ; <float> [#uses=1]
  %32 = getelementptr float* %a0, i32 2           ; <float*> [#uses=1]
  %pos.z.a0 = load float* %32                     ; <float> [#uses=1]
  %33 = insertelement <4 x float> <float 0.000000e+00, float undef, float undef, float undef>, float %pos.z.dadx, i32 1 ; <<4 x float>> [#uses=1]
  %34 = insertelement <4 x float> %33, float %pos.z.dady, i32 2 ; <<4 x float>> [#uses=1]
  %pos.z.dadq = insertelement <4 x float> %34, float %pos.z.dadxy, i32 3 ; <<4 x float>> [#uses=6]
  %35 = add <4 x float> %pos.z.dadq, %pos.z.dadq  ; <<4 x float>> [#uses=1]
  %36 = mul float %18, %pos.z.dadx                ; <float> [#uses=1]
  %37 = add float %pos.z.a0, %36                  ; <float> [#uses=1]
  %38 = mul float %19, %pos.z.dady                ; <float> [#uses=1]
  %39 = add float %37, %38                        ; <float> [#uses=4]
  %40 = insertelement <4 x float> undef, float %39, i32 0 ; <<4 x float>> [#uses=1]
  %41 = insertelement <4 x float> %40, float %39, i32 1 ; <<4 x float>> [#uses=1]
  %42 = insertelement <4 x float> %41, float %39, i32 2 ; <<4 x float>> [#uses=1]
  %43 = insertelement <4 x float> %42, float %39, i32 3 ; <<4 x float>> [#uses=1]
  %pos.z.a = add <4 x float> %43, %35             ; <<4 x float>> [#uses=4]
  %44 = getelementptr float* %dadx, i32 3         ; <float*> [#uses=1]
  %pos.w.dadx = load float* %44                   ; <float> [#uses=3]
  %45 = getelementptr float* %dady, i32 3         ; <float*> [#uses=1]
  %pos.w.dady = load float* %45                   ; <float> [#uses=3]
  %pos.w.dadxy = add float %pos.w.dadx, %pos.w.dady ; <float> [#uses=1]
  %46 = getelementptr float* %a0, i32 3           ; <float*> [#uses=1]
  %pos.w.a0 = load float* %46                     ; <float> [#uses=1]
  %47 = insertelement <4 x float> <float 0.000000e+00, float undef, float undef, float undef>, float %pos.w.dadx, i32 1 ; <<4 x float>> [#uses=1]
  %48 = insertelement <4 x float> %47, float %pos.w.dady, i32 2 ; <<4 x float>> [#uses=1]
  %pos.w.dadq = insertelement <4 x float> %48, float %pos.w.dadxy, i32 3 ; <<4 x float>> [#uses=6]
  %49 = add <4 x float> %pos.w.dadq, %pos.w.dadq  ; <<4 x float>> [#uses=1]
  %50 = mul float %18, %pos.w.dadx                ; <float> [#uses=1]
  %51 = add float %pos.w.a0, %50                  ; <float> [#uses=1]
  %52 = mul float %19, %pos.w.dady                ; <float> [#uses=1]
  %53 = add float %51, %52                        ; <float> [#uses=4]
  %54 = insertelement <4 x float> undef, float %53, i32 0 ; <<4 x float>> [#uses=1]
  %55 = insertelement <4 x float> %54, float %53, i32 1 ; <<4 x float>> [#uses=1]
  %56 = insertelement <4 x float> %55, float %53, i32 2 ; <<4 x float>> [#uses=1]
  %57 = insertelement <4 x float> %56, float %53, i32 3 ; <<4 x float>> [#uses=1]
  %pos.w.a = add <4 x float> %57, %49             ; <<4 x float>> [#uses=4]
  %58 = getelementptr float* %a0, i32 4           ; <float*> [#uses=1]
  %input0.x.a0 = load float* %58                  ; <float> [#uses=4]
  %59 = insertelement <4 x float> undef, float %input0.x.a0, i32 0 ; <<4 x float>> [#uses=1]
  %60 = insertelement <4 x float> %59, float %input0.x.a0, i32 1 ; <<4 x float>> [#uses=1]
  %61 = insertelement <4 x float> %60, float %input0.x.a0, i32 2 ; <<4 x float>> [#uses=1]
  %62 = insertelement <4 x float> %61, float %input0.x.a0, i32 3 ; <<4 x float>> [#uses=1]
  %input0.x.a = add <4 x float> %62, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
  %63 = getelementptr float* %a0, i32 5           ; <float*> [#uses=1]
  %input0.y.a0 = load float* %63                  ; <float> [#uses=4]
  %64 = insertelement <4 x float> undef, float %input0.y.a0, i32 0 ; <<4 x float>> [#uses=1]
  %65 = insertelement <4 x float> %64, float %input0.y.a0, i32 1 ; <<4 x float>> [#uses=1]
  %66 = insertelement <4 x float> %65, float %input0.y.a0, i32 2 ; <<4 x float>> [#uses=1]
  %67 = insertelement <4 x float> %66, float %input0.y.a0, i32 3 ; <<4 x float>> [#uses=1]
  %input0.y.a = add <4 x float> %67, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
  %68 = getelementptr float* %a0, i32 6           ; <float*> [#uses=1]
  %input0.z.a0 = load float* %68                  ; <float> [#uses=4]
  %69 = insertelement <4 x float> undef, float %input0.z.a0, i32 0 ; <<4 x float>> [#uses=1]
  %70 = insertelement <4 x float> %69, float %input0.z.a0, i32 1 ; <<4 x float>> [#uses=1]
  %71 = insertelement <4 x float> %70, float %input0.z.a0, i32 2 ; <<4 x float>> [#uses=1]
  %72 = insertelement <4 x float> %71, float %input0.z.a0, i32 3 ; <<4 x float>> [#uses=1]
  %input0.z.a = add <4 x float> %72, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
  %73 = getelementptr float* %a0, i32 7           ; <float*> [#uses=1]
  %input0.w.a0 = load float* %73                  ; <float> [#uses=4]
  %74 = insertelement <4 x float> undef, float %input0.w.a0, i32 0 ; <<4 x float>> [#uses=1]
  %75 = insertelement <4 x float> %74, float %input0.w.a0, i32 1 ; <<4 x float>> [#uses=1]
  %76 = insertelement <4 x float> %75, float %input0.w.a0, i32 2 ; <<4 x float>> [#uses=1]
  %77 = insertelement <4 x float> %76, float %input0.w.a0, i32 3 ; <<4 x float>> [#uses=1]
  %input0.w.a = add <4 x float> %77, <float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00), float add (float 0.000000e+00, float 0.000000e+00)> ; <<4 x float>> [#uses=4]
  %78 = shufflevector <4 x float> %pos.x.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %pos.x = add <4 x float> %78, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %79 = shufflevector <4 x float> %pos.y.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %pos.y = add <4 x float> %79, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %80 = shufflevector <4 x float> %pos.z.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %pos.z = add <4 x float> %80, %pos.z.dadq       ; <<4 x float>> [#uses=4]
  %81 = shufflevector <4 x float> %pos.w.a, <4 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
  %pos.w = add <4 x float> %81, %pos.w.dadq       ; <<4 x float>> [#uses=0]
  %82 = getelementptr <4 x i32>* %depth, i32 0    ; <<4 x i32>*> [#uses=2]
  %context.stencil_ref_front_ptr = getelementptr %context* %context, i32 0, i32 2 ; <i32*> [#uses=1]
  %context.stencil_ref_front = load i32* %context.stencil_ref_front_ptr ; <i32> [#uses=0]
  %context.stencil_ref_back_ptr = getelementptr %context* %context, i32 0, i32 3 ; <i32*> [#uses=1]
  %context.stencil_ref_back = load i32* %context.stencil_ref_back_ptr ; <i32> [#uses=0]
  %context.constants_ptr = getelementptr %context* %context, i32 0, i32 0 ; <float**> [#uses=1]
  %context.constants = load float** %context.constants_ptr ; <float*> [#uses=0]
  br i1 icmp eq (i128 bitcast (<4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1> to i128), i128 0), label %106, label %83

; <label>:83                                      ; preds = %entry
  %84 = mul <4 x float> %pos.z, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
  %85 = add <4 x float> %84, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %86 = bitcast <4 x float> %85 to <4 x i32>      ; <<4 x i32>> [#uses=1]
  %87 = shl <4 x i32> %86, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %zsbufval = load <4 x i32>* %82                 ; <<4 x i32>> [#uses=1]
  %88 = lshr <4 x i32> %87, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %89 = and <4 x i32> %88, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %zsbuf.z = and <4 x i32> %zsbufval, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %90 = icmp ult <4 x i32> %89, %zsbuf.z          ; <<4 x i1>> [#uses=1]
  %91 = sext <4 x i1> %90 to <4 x i32>            ; <<4 x i32>> [#uses=2]
  %92 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %91 ; <<4 x i32>> [#uses=4]
  %93 = bitcast <4 x i32> %92 to i128             ; <i128> [#uses=1]
  %94 = icmp eq i128 %93, 0                       ; <i1> [#uses=1]
  br i1 %94, label %106, label %95

; <label>:95                                      ; preds = %83
  %96 = and <4 x i32> %92, %91                    ; <<4 x i32>> [#uses=1]
  %97 = and <4 x i32> %96, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %98 = and <4 x i32> %89, %97                    ; <<4 x i32>> [#uses=1]
  %99 = xor <4 x i32> %97, <i32 -1, i32 -1, i32 -1, i32 -1> ; <<4 x i32>> [#uses=1]
  %100 = and <4 x i32> %zsbuf.z, %99              ; <<4 x i32>> [#uses=1]
  %101 = or <4 x i32> %98, %100                   ; <<4 x i32>> [#uses=1]
  store <4 x i32> %101, <4 x i32>* %82
  store <4 x float> %input0.x.a, <4 x float>* %17
  store <4 x float> %input0.y.a, <4 x float>* %16
  store <4 x float> %input0.z.a, <4 x float>* %15
  store <4 x float> %input0.w.a, <4 x float>* %14
  %color0.0.r = load <4 x float>* %17             ; <<4 x float>> [#uses=1]
  %color0.0.g = load <4 x float>* %16             ; <<4 x float>> [#uses=1]
  %color0.0.b = load <4 x float>* %15             ; <<4 x float>> [#uses=1]
  %color0.0.a = load <4 x float>* %14             ; <<4 x float>> [#uses=1]
  %context.alpha_ref_value_ptr = getelementptr %context* %context, i32 0, i32 1 ; <float*> [#uses=1]
  %context.alpha_ref_value = load float* %context.alpha_ref_value_ptr ; <float> [#uses=4]
  %102 = insertelement <4 x float> undef, float %context.alpha_ref_value, i32 0 ; <<4 x float>> [#uses=1]
  %103 = insertelement <4 x float> %102, float %context.alpha_ref_value, i32 1 ; <<4 x float>> [#uses=1]
  %104 = insertelement <4 x float> %103, float %context.alpha_ref_value, i32 2 ; <<4 x float>> [#uses=1]
  %105 = insertelement <4 x float> %104, float %context.alpha_ref_value, i32 3 ; <<4 x float>> [#uses=0]
  br label %106

; <label>:106                                     ; preds = %95, %83, %entry
  %107 = phi <4 x float> [ undef, %entry ], [ undef, %83 ], [ %color0.0.r, %95 ] ; <<4 x float>> [#uses=1]
  %108 = phi <4 x float> [ undef, %entry ], [ undef, %83 ], [ %color0.0.g, %95 ] ; <<4 x float>> [#uses=1]
  %109 = phi <4 x float> [ undef, %entry ], [ undef, %83 ], [ %color0.0.b, %95 ] ; <<4 x float>> [#uses=1]
  %110 = phi <4 x float> [ undef, %entry ], [ undef, %83 ], [ %color0.0.a, %95 ] ; <<4 x float>> [#uses=1]
  %111 = phi <4 x float> [ %pos.z, %entry ], [ %pos.z, %83 ], [ %pos.z, %95 ] ; <<4 x float>> [#uses=0]
  %112 = phi <4 x i32> [ <i32 -1, i32 -1, i32 -1, i32 -1>, %entry ], [ %92, %83 ], [ %92, %95 ] ; <<4 x i32>> [#uses=1]
  %113 = shufflevector <4 x float> %pos.x.a, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x float>> [#uses=1]
  %pos.x1 = add <4 x float> %113, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %114 = shufflevector <4 x float> %pos.y.a, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x float>> [#uses=1]
  %pos.y2 = add <4 x float> %114, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %115 = shufflevector <4 x float> %pos.z.a, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x float>> [#uses=1]
  %pos.z3 = add <4 x float> %115, %pos.z.dadq     ; <<4 x float>> [#uses=4]
  %116 = shufflevector <4 x float> %pos.w.a, <4 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x float>> [#uses=1]
  %pos.w4 = add <4 x float> %116, %pos.w.dadq     ; <<4 x float>> [#uses=0]
  %117 = getelementptr <4 x i32>* %depth, i32 1   ; <<4 x i32>*> [#uses=2]
  %context.stencil_ref_front_ptr5 = getelementptr %context* %context, i32 0, i32 2 ; <i32*> [#uses=1]
  %context.stencil_ref_front6 = load i32* %context.stencil_ref_front_ptr5 ; <i32> [#uses=0]
  %context.stencil_ref_back_ptr7 = getelementptr %context* %context, i32 0, i32 3 ; <i32*> [#uses=1]
  %context.stencil_ref_back8 = load i32* %context.stencil_ref_back_ptr7 ; <i32> [#uses=0]
  %context.constants_ptr9 = getelementptr %context* %context, i32 0, i32 0 ; <float**> [#uses=1]
  %context.constants10 = load float** %context.constants_ptr9 ; <float*> [#uses=0]
  br i1 icmp eq (i128 bitcast (<4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1> to i128), i128 0), label %141, label %118

; <label>:118                                     ; preds = %106
  %119 = mul <4 x float> %pos.z3, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
  %120 = add <4 x float> %119, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %121 = bitcast <4 x float> %120 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %122 = shl <4 x i32> %121, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %zsbufval11 = load <4 x i32>* %117              ; <<4 x i32>> [#uses=1]
  %123 = lshr <4 x i32> %122, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %124 = and <4 x i32> %123, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %zsbuf.z12 = and <4 x i32> %zsbufval11, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %125 = icmp ult <4 x i32> %124, %zsbuf.z12      ; <<4 x i1>> [#uses=1]
  %126 = sext <4 x i1> %125 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %127 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %126 ; <<4 x i32>> [#uses=4]
  %128 = bitcast <4 x i32> %127 to i128           ; <i128> [#uses=1]
  %129 = icmp eq i128 %128, 0                     ; <i1> [#uses=1]
  br i1 %129, label %141, label %130

; <label>:130                                     ; preds = %118
  %131 = and <4 x i32> %127, %126                 ; <<4 x i32>> [#uses=1]
  %132 = and <4 x i32> %131, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %133 = and <4 x i32> %124, %132                 ; <<4 x i32>> [#uses=1]
  %134 = xor <4 x i32> %132, <i32 -1, i32 -1, i32 -1, i32 -1> ; <<4 x i32>> [#uses=1]
  %135 = and <4 x i32> %zsbuf.z12, %134           ; <<4 x i32>> [#uses=1]
  %136 = or <4 x i32> %133, %135                  ; <<4 x i32>> [#uses=1]
  store <4 x i32> %136, <4 x i32>* %117
  store <4 x float> %input0.x.a, <4 x float>* %13
  store <4 x float> %input0.y.a, <4 x float>* %12
  store <4 x float> %input0.z.a, <4 x float>* %11
  store <4 x float> %input0.w.a, <4 x float>* %10
  %color1.0.r = load <4 x float>* %13             ; <<4 x float>> [#uses=1]
  %color1.0.g = load <4 x float>* %12             ; <<4 x float>> [#uses=1]
  %color1.0.b = load <4 x float>* %11             ; <<4 x float>> [#uses=1]
  %color1.0.a = load <4 x float>* %10             ; <<4 x float>> [#uses=1]
  %context.alpha_ref_value_ptr13 = getelementptr %context* %context, i32 0, i32 1 ; <float*> [#uses=1]
  %context.alpha_ref_value14 = load float* %context.alpha_ref_value_ptr13 ; <float> [#uses=4]
  %137 = insertelement <4 x float> undef, float %context.alpha_ref_value14, i32 0 ; <<4 x float>> [#uses=1]
  %138 = insertelement <4 x float> %137, float %context.alpha_ref_value14, i32 1 ; <<4 x float>> [#uses=1]
  %139 = insertelement <4 x float> %138, float %context.alpha_ref_value14, i32 2 ; <<4 x float>> [#uses=1]
  %140 = insertelement <4 x float> %139, float %context.alpha_ref_value14, i32 3 ; <<4 x float>> [#uses=0]
  br label %141

; <label>:141                                     ; preds = %130, %118, %106
  %142 = phi <4 x float> [ undef, %106 ], [ undef, %118 ], [ %color1.0.r, %130 ] ; <<4 x float>> [#uses=1]
  %143 = phi <4 x float> [ undef, %106 ], [ undef, %118 ], [ %color1.0.g, %130 ] ; <<4 x float>> [#uses=1]
  %144 = phi <4 x float> [ undef, %106 ], [ undef, %118 ], [ %color1.0.b, %130 ] ; <<4 x float>> [#uses=1]
  %145 = phi <4 x float> [ undef, %106 ], [ undef, %118 ], [ %color1.0.a, %130 ] ; <<4 x float>> [#uses=1]
  %146 = phi <4 x float> [ %pos.z3, %106 ], [ %pos.z3, %118 ], [ %pos.z3, %130 ] ; <<4 x float>> [#uses=0]
  %147 = phi <4 x i32> [ <i32 -1, i32 -1, i32 -1, i32 -1>, %106 ], [ %127, %118 ], [ %127, %130 ] ; <<4 x i32>> [#uses=1]
  %148 = shufflevector <4 x float> %pos.x.a, <4 x float> undef, <4 x i32> <i32 2, i32 2, i32 2, i32 2> ; <<4 x float>> [#uses=1]
  %pos.x15 = add <4 x float> %148, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %149 = shufflevector <4 x float> %pos.y.a, <4 x float> undef, <4 x i32> <i32 2, i32 2, i32 2, i32 2> ; <<4 x float>> [#uses=1]
  %pos.y16 = add <4 x float> %149, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %150 = shufflevector <4 x float> %pos.z.a, <4 x float> undef, <4 x i32> <i32 2, i32 2, i32 2, i32 2> ; <<4 x float>> [#uses=1]
  %pos.z17 = add <4 x float> %150, %pos.z.dadq    ; <<4 x float>> [#uses=4]
  %151 = shufflevector <4 x float> %pos.w.a, <4 x float> undef, <4 x i32> <i32 2, i32 2, i32 2, i32 2> ; <<4 x float>> [#uses=1]
  %pos.w18 = add <4 x float> %151, %pos.w.dadq    ; <<4 x float>> [#uses=0]
  %152 = getelementptr <4 x i32>* %depth, i32 2   ; <<4 x i32>*> [#uses=2]
  %context.stencil_ref_front_ptr19 = getelementptr %context* %context, i32 0, i32 2 ; <i32*> [#uses=1]
  %context.stencil_ref_front20 = load i32* %context.stencil_ref_front_ptr19 ; <i32> [#uses=0]
  %context.stencil_ref_back_ptr21 = getelementptr %context* %context, i32 0, i32 3 ; <i32*> [#uses=1]
  %context.stencil_ref_back22 = load i32* %context.stencil_ref_back_ptr21 ; <i32> [#uses=0]
  %context.constants_ptr23 = getelementptr %context* %context, i32 0, i32 0 ; <float**> [#uses=1]
  %context.constants24 = load float** %context.constants_ptr23 ; <float*> [#uses=0]
  br i1 icmp eq (i128 bitcast (<4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1> to i128), i128 0), label %176, label %153

; <label>:153                                     ; preds = %141
  %154 = mul <4 x float> %pos.z17, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
  %155 = add <4 x float> %154, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %156 = bitcast <4 x float> %155 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %157 = shl <4 x i32> %156, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %zsbufval25 = load <4 x i32>* %152              ; <<4 x i32>> [#uses=1]
  %158 = lshr <4 x i32> %157, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %159 = and <4 x i32> %158, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %zsbuf.z26 = and <4 x i32> %zsbufval25, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %160 = icmp ult <4 x i32> %159, %zsbuf.z26      ; <<4 x i1>> [#uses=1]
  %161 = sext <4 x i1> %160 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %162 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %161 ; <<4 x i32>> [#uses=4]
  %163 = bitcast <4 x i32> %162 to i128           ; <i128> [#uses=1]
  %164 = icmp eq i128 %163, 0                     ; <i1> [#uses=1]
  br i1 %164, label %176, label %165

; <label>:165                                     ; preds = %153
  %166 = and <4 x i32> %162, %161                 ; <<4 x i32>> [#uses=1]
  %167 = and <4 x i32> %166, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %168 = and <4 x i32> %159, %167                 ; <<4 x i32>> [#uses=1]
  %169 = xor <4 x i32> %167, <i32 -1, i32 -1, i32 -1, i32 -1> ; <<4 x i32>> [#uses=1]
  %170 = and <4 x i32> %zsbuf.z26, %169           ; <<4 x i32>> [#uses=1]
  %171 = or <4 x i32> %168, %170                  ; <<4 x i32>> [#uses=1]
  store <4 x i32> %171, <4 x i32>* %152
  store <4 x float> %input0.x.a, <4 x float>* %9
  store <4 x float> %input0.y.a, <4 x float>* %8
  store <4 x float> %input0.z.a, <4 x float>* %7
  store <4 x float> %input0.w.a, <4 x float>* %6
  %color2.0.r = load <4 x float>* %9              ; <<4 x float>> [#uses=1]
  %color2.0.g = load <4 x float>* %8              ; <<4 x float>> [#uses=1]
  %color2.0.b = load <4 x float>* %7              ; <<4 x float>> [#uses=1]
  %color2.0.a = load <4 x float>* %6              ; <<4 x float>> [#uses=1]
  %context.alpha_ref_value_ptr27 = getelementptr %context* %context, i32 0, i32 1 ; <float*> [#uses=1]
  %context.alpha_ref_value28 = load float* %context.alpha_ref_value_ptr27 ; <float> [#uses=4]
  %172 = insertelement <4 x float> undef, float %context.alpha_ref_value28, i32 0 ; <<4 x float>> [#uses=1]
  %173 = insertelement <4 x float> %172, float %context.alpha_ref_value28, i32 1 ; <<4 x float>> [#uses=1]
  %174 = insertelement <4 x float> %173, float %context.alpha_ref_value28, i32 2 ; <<4 x float>> [#uses=1]
  %175 = insertelement <4 x float> %174, float %context.alpha_ref_value28, i32 3 ; <<4 x float>> [#uses=0]
  br label %176

; <label>:176                                     ; preds = %165, %153, %141
  %177 = phi <4 x float> [ undef, %141 ], [ undef, %153 ], [ %color2.0.r, %165 ] ; <<4 x float>> [#uses=1]
  %178 = phi <4 x float> [ undef, %141 ], [ undef, %153 ], [ %color2.0.g, %165 ] ; <<4 x float>> [#uses=1]
  %179 = phi <4 x float> [ undef, %141 ], [ undef, %153 ], [ %color2.0.b, %165 ] ; <<4 x float>> [#uses=1]
  %180 = phi <4 x float> [ undef, %141 ], [ undef, %153 ], [ %color2.0.a, %165 ] ; <<4 x float>> [#uses=1]
  %181 = phi <4 x float> [ %pos.z17, %141 ], [ %pos.z17, %153 ], [ %pos.z17, %165 ] ; <<4 x float>> [#uses=0]
  %182 = phi <4 x i32> [ <i32 -1, i32 -1, i32 -1, i32 -1>, %141 ], [ %162, %153 ], [ %162, %165 ] ; <<4 x i32>> [#uses=1]
  %183 = shufflevector <4 x float> %pos.x.a, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3> ; <<4 x float>> [#uses=1]
  %pos.x29 = add <4 x float> %183, <float 0.000000e+00, float 1.000000e+00, float 0.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %184 = shufflevector <4 x float> %pos.y.a, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3> ; <<4 x float>> [#uses=1]
  %pos.y30 = add <4 x float> %184, <float 0.000000e+00, float 0.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=0]
  %185 = shufflevector <4 x float> %pos.z.a, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3> ; <<4 x float>> [#uses=1]
  %pos.z31 = add <4 x float> %185, %pos.z.dadq    ; <<4 x float>> [#uses=4]
  %186 = shufflevector <4 x float> %pos.w.a, <4 x float> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3> ; <<4 x float>> [#uses=1]
  %pos.w32 = add <4 x float> %186, %pos.w.dadq    ; <<4 x float>> [#uses=0]
  %187 = getelementptr <4 x i32>* %depth, i32 3   ; <<4 x i32>*> [#uses=2]
  %context.stencil_ref_front_ptr33 = getelementptr %context* %context, i32 0, i32 2 ; <i32*> [#uses=1]
  %context.stencil_ref_front34 = load i32* %context.stencil_ref_front_ptr33 ; <i32> [#uses=0]
  %context.stencil_ref_back_ptr35 = getelementptr %context* %context, i32 0, i32 3 ; <i32*> [#uses=1]
  %context.stencil_ref_back36 = load i32* %context.stencil_ref_back_ptr35 ; <i32> [#uses=0]
  %context.constants_ptr37 = getelementptr %context* %context, i32 0, i32 0 ; <float**> [#uses=1]
  %context.constants38 = load float** %context.constants_ptr37 ; <float*> [#uses=0]
  br i1 icmp eq (i128 bitcast (<4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1> to i128), i128 0), label %211, label %188

; <label>:188                                     ; preds = %176
  %189 = mul <4 x float> %pos.z31, <float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000, float 0x3FEFFFFFC0000000> ; <<4 x float>> [#uses=1]
  %190 = add <4 x float> %189, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00> ; <<4 x float>> [#uses=1]
  %191 = bitcast <4 x float> %190 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %192 = shl <4 x i32> %191, <i32 9, i32 9, i32 9, i32 9> ; <<4 x i32>> [#uses=1]
  %zsbufval39 = load <4 x i32>* %187              ; <<4 x i32>> [#uses=1]
  %193 = lshr <4 x i32> %192, <i32 8, i32 8, i32 8, i32 8> ; <<4 x i32>> [#uses=1]
  %194 = and <4 x i32> %193, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %zsbuf.z40 = and <4 x i32> %zsbufval39, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %195 = icmp ult <4 x i32> %194, %zsbuf.z40      ; <<4 x i1>> [#uses=1]
  %196 = sext <4 x i1> %195 to <4 x i32>          ; <<4 x i32>> [#uses=2]
  %197 = and <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>, %196 ; <<4 x i32>> [#uses=4]
  %198 = bitcast <4 x i32> %197 to i128           ; <i128> [#uses=1]
  %199 = icmp eq i128 %198, 0                     ; <i1> [#uses=1]
  br i1 %199, label %211, label %200

; <label>:200                                     ; preds = %188
  %201 = and <4 x i32> %197, %196                 ; <<4 x i32>> [#uses=1]
  %202 = and <4 x i32> %201, <i32 16777215, i32 16777215, i32 16777215, i32 16777215> ; <<4 x i32>> [#uses=2]
  %203 = and <4 x i32> %194, %202                 ; <<4 x i32>> [#uses=1]
  %204 = xor <4 x i32> %202, <i32 -1, i32 -1, i32 -1, i32 -1> ; <<4 x i32>> [#uses=1]
  %205 = and <4 x i32> %zsbuf.z40, %204           ; <<4 x i32>> [#uses=1]
  %206 = or <4 x i32> %203, %205                  ; <<4 x i32>> [#uses=1]
  store <4 x i32> %206, <4 x i32>* %187
  store <4 x float> %input0.x.a, <4 x float>* %5
  store <4 x float> %input0.y.a, <4 x float>* %4
  store <4 x float> %input0.z.a, <4 x float>* %3
  store <4 x float> %input0.w.a, <4 x float>* %2
  %color3.0.r = load <4 x float>* %5              ; <<4 x float>> [#uses=1]
  %color3.0.g = load <4 x float>* %4              ; <<4 x float>> [#uses=1]
  %color3.0.b = load <4 x float>* %3              ; <<4 x float>> [#uses=1]
  %color3.0.a = load <4 x float>* %2              ; <<4 x float>> [#uses=1]
  %context.alpha_ref_value_ptr41 = getelementptr %context* %context, i32 0, i32 1 ; <float*> [#uses=1]
  %context.alpha_ref_value42 = load float* %context.alpha_ref_value_ptr41 ; <float> [#uses=4]
  %207 = insertelement <4 x float> undef, float %context.alpha_ref_value42, i32 0 ; <<4 x float>> [#uses=1]
  %208 = insertelement <4 x float> %207, float %context.alpha_ref_value42, i32 1 ; <<4 x float>> [#uses=1]
  %209 = insertelement <4 x float> %208, float %context.alpha_ref_value42, i32 2 ; <<4 x float>> [#uses=1]
  %210 = insertelement <4 x float> %209, float %context.alpha_ref_value42, i32 3 ; <<4 x float>> [#uses=0]
  br label %211

; <label>:211                                     ; preds = %200, %188, %176
  %212 = phi <4 x float> [ undef, %176 ], [ undef, %188 ], [ %color3.0.r, %200 ] ; <<4 x float>> [#uses=1]
  %213 = phi <4 x float> [ undef, %176 ], [ undef, %188 ], [ %color3.0.g, %200 ] ; <<4 x float>> [#uses=1]
  %214 = phi <4 x float> [ undef, %176 ], [ undef, %188 ], [ %color3.0.b, %200 ] ; <<4 x float>> [#uses=1]
  %215 = phi <4 x float> [ undef, %176 ], [ undef, %188 ], [ %color3.0.a, %200 ] ; <<4 x float>> [#uses=1]
  %216 = phi <4 x float> [ %pos.z31, %176 ], [ %pos.z31, %188 ], [ %pos.z31, %200 ] ; <<4 x float>> [#uses=0]
  %217 = phi <4 x i32> [ <i32 -1, i32 -1, i32 -1, i32 -1>, %176 ], [ %197, %188 ], [ %197, %200 ] ; <<4 x i32>> [#uses=1]
  %218 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %107, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %219 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %142, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %220 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %177, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %221 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %212, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %222 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %218, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %223 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %219, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %224 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %220, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %225 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %221, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %226 = mul <4 x float> %222, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %227 = add <4 x float> %226, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %228 = bitcast <4 x float> %227 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %229 = and <4 x i32> %228, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %230 = mul <4 x float> %223, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %231 = add <4 x float> %230, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %232 = bitcast <4 x float> %231 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %233 = and <4 x i32> %232, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %234 = mul <4 x float> %224, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %235 = add <4 x float> %234, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %236 = bitcast <4 x float> %235 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %237 = and <4 x i32> %236, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %238 = mul <4 x float> %225, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %239 = add <4 x float> %238, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %240 = bitcast <4 x float> %239 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %241 = and <4 x i32> %240, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %242 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %229, <4 x i32> %233) ; <<8 x i16>> [#uses=1]
  %243 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %237, <4 x i32> %241) ; <<8 x i16>> [#uses=1]
  %res.r = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %242, <8 x i16> %243) ; <<16 x i8>> [#uses=1]
  %244 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %108, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %245 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %143, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %246 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %178, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %247 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %213, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %248 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %244, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %249 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %245, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %250 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %246, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %251 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %247, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %252 = mul <4 x float> %248, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %253 = add <4 x float> %252, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %254 = bitcast <4 x float> %253 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %255 = and <4 x i32> %254, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %256 = mul <4 x float> %249, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %257 = add <4 x float> %256, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %258 = bitcast <4 x float> %257 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %259 = and <4 x i32> %258, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %260 = mul <4 x float> %250, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %261 = add <4 x float> %260, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %262 = bitcast <4 x float> %261 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %263 = and <4 x i32> %262, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %264 = mul <4 x float> %251, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %265 = add <4 x float> %264, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %266 = bitcast <4 x float> %265 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %267 = and <4 x i32> %266, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %268 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %255, <4 x i32> %259) ; <<8 x i16>> [#uses=1]
  %269 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %263, <4 x i32> %267) ; <<8 x i16>> [#uses=1]
  %res.g = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %268, <8 x i16> %269) ; <<16 x i8>> [#uses=1]
  %270 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %109, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %271 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %144, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %272 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %179, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %273 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %214, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %274 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %270, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %275 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %271, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %276 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %272, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %277 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %273, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %278 = mul <4 x float> %274, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %279 = add <4 x float> %278, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %280 = bitcast <4 x float> %279 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %281 = and <4 x i32> %280, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %282 = mul <4 x float> %275, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %283 = add <4 x float> %282, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %284 = bitcast <4 x float> %283 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %285 = and <4 x i32> %284, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %286 = mul <4 x float> %276, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %287 = add <4 x float> %286, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %288 = bitcast <4 x float> %287 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %289 = and <4 x i32> %288, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %290 = mul <4 x float> %277, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %291 = add <4 x float> %290, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %292 = bitcast <4 x float> %291 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %293 = and <4 x i32> %292, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %294 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %281, <4 x i32> %285) ; <<8 x i16>> [#uses=1]
  %295 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %289, <4 x i32> %293) ; <<8 x i16>> [#uses=1]
  %res.b = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %294, <8 x i16> %295) ; <<16 x i8>> [#uses=1]
  %296 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %110, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %297 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %145, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %298 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %180, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %299 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %215, <4 x float> zeroinitializer) ; <<4 x float>> [#uses=1]
  %300 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %296, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %301 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %297, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %302 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %298, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %303 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %299, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>) ; <<4 x float>> [#uses=1]
  %304 = mul <4 x float> %300, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %305 = add <4 x float> %304, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %306 = bitcast <4 x float> %305 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %307 = and <4 x i32> %306, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %308 = mul <4 x float> %301, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %309 = add <4 x float> %308, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %310 = bitcast <4 x float> %309 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %311 = and <4 x i32> %310, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %312 = mul <4 x float> %302, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %313 = add <4 x float> %312, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %314 = bitcast <4 x float> %313 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %315 = and <4 x i32> %314, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %316 = mul <4 x float> %303, <float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000, float 0x3FEFE00000000000> ; <<4 x float>> [#uses=1]
  %317 = add <4 x float> %316, <float 3.276800e+04, float 3.276800e+04, float 3.276800e+04, float 3.276800e+04> ; <<4 x float>> [#uses=1]
  %318 = bitcast <4 x float> %317 to <4 x i32>    ; <<4 x i32>> [#uses=1]
  %319 = and <4 x i32> %318, <i32 255, i32 255, i32 255, i32 255> ; <<4 x i32>> [#uses=1]
  %320 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %307, <4 x i32> %311) ; <<8 x i16>> [#uses=1]
  %321 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %315, <4 x i32> %319) ; <<8 x i16>> [#uses=1]
  %color0.a = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %320, <8 x i16> %321) ; <<16 x i8>> [#uses=0]
  %322 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %112, <4 x i32> %147) ; <<8 x i16>> [#uses=1]
  %323 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %182, <4 x i32> %217) ; <<8 x i16>> [#uses=1]
  %324 = call <16 x i8> @llvm.x86.sse2.packsswb.128(<8 x i16> %322, <8 x i16> %323) ; <<16 x i8>> [#uses=9]
  %325 = getelementptr <16 x i8>** %color_ptr_ptr, i32 0 ; <<16 x i8>**> [#uses=1]
  %color_ptr0 = load <16 x i8>** %325             ; <<16 x i8>*> [#uses=7]
  %326 = bitcast <16 x i8> %324 to i128           ; <i128> [#uses=1]
  %327 = icmp eq i128 %326, 0                     ; <i1> [#uses=1]
  br i1 %327, label %353, label %328

; <label>:328                                     ; preds = %211
  %context.blend_color_ptr = getelementptr %context* %context, i32 0, i32 4 ; <i8**> [#uses=1]
  %context.blend_color = load i8** %context.blend_color_ptr ; <i8*> [#uses=1]
  %329 = bitcast i8* %context.blend_color to <16 x i8>* ; <<16 x i8>*> [#uses=4]
  %330 = getelementptr <16 x i8>* %329, i32 0     ; <<16 x i8>*> [#uses=1]
  %con.r = load <16 x i8>* %330                   ; <<16 x i8>> [#uses=0]
  %331 = getelementptr <16 x i8>* %color_ptr0, i32 0 ; <<16 x i8>*> [#uses=1]
  %dst.r = load <16 x i8>* %331                   ; <<16 x i8>> [#uses=1]
  %332 = getelementptr <16 x i8>* %329, i32 1     ; <<16 x i8>*> [#uses=1]
  %con.g = load <16 x i8>* %332                   ; <<16 x i8>> [#uses=0]
  %333 = getelementptr <16 x i8>* %color_ptr0, i32 1 ; <<16 x i8>*> [#uses=1]
  %dst.g = load <16 x i8>* %333                   ; <<16 x i8>> [#uses=1]
  %334 = getelementptr <16 x i8>* %329, i32 2     ; <<16 x i8>*> [#uses=1]
  %con.b = load <16 x i8>* %334                   ; <<16 x i8>> [#uses=0]
  %335 = getelementptr <16 x i8>* %color_ptr0, i32 2 ; <<16 x i8>*> [#uses=1]
  %dst.b = load <16 x i8>* %335                   ; <<16 x i8>> [#uses=1]
  %336 = getelementptr <16 x i8>* %329, i32 3     ; <<16 x i8>*> [#uses=1]
  %con.a = load <16 x i8>* %336                   ; <<16 x i8>> [#uses=0]
  %337 = getelementptr <16 x i8>* %color_ptr0, i32 3 ; <<16 x i8>*> [#uses=1]
  %dst.a = load <16 x i8>* %337                   ; <<16 x i8>> [#uses=0]
  %338 = and <16 x i8> %res.r, %324               ; <<16 x i8>> [#uses=1]
  %339 = xor <16 x i8> %324, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> ; <<16 x i8>> [#uses=1]
  %340 = and <16 x i8> %dst.r, %339               ; <<16 x i8>> [#uses=1]
  %341 = or <16 x i8> %338, %340                  ; <<16 x i8>> [#uses=1]
  %342 = getelementptr <16 x i8>* %color_ptr0, i32 0 ; <<16 x i8>*> [#uses=1]
  store <16 x i8> %341, <16 x i8>* %342
  %343 = and <16 x i8> %res.g, %324               ; <<16 x i8>> [#uses=1]
  %344 = xor <16 x i8> %324, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> ; <<16 x i8>> [#uses=1]
  %345 = and <16 x i8> %dst.g, %344               ; <<16 x i8>> [#uses=1]
  %346 = or <16 x i8> %343, %345                  ; <<16 x i8>> [#uses=1]
  %347 = getelementptr <16 x i8>* %color_ptr0, i32 1 ; <<16 x i8>*> [#uses=1]
  store <16 x i8> %346, <16 x i8>* %347
  %348 = and <16 x i8> %res.b, %324               ; <<16 x i8>> [#uses=1]
  %349 = xor <16 x i8> %324, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> ; <<16 x i8>> [#uses=1]
  %350 = and <16 x i8> %dst.b, %349               ; <<16 x i8>> [#uses=1]
  %351 = or <16 x i8> %348, %350                  ; <<16 x i8>> [#uses=1]
  %352 = getelementptr <16 x i8>* %color_ptr0, i32 2 ; <<16 x i8>*> [#uses=1]
  store <16 x i8> %351, <16 x i8>* %352
  br label %353

; <label>:353                                     ; preds = %328, %211
  %354 = phi <16 x i8> [ %324, %211 ], [ %324, %328 ] ; <<16 x i8>> [#uses=0]
  ret void
}