Dan Gohman | 1880092 | 2009-09-11 18:01:28 +0000 | [diff] [blame] | 1 | ; RUN: opt < %s -instcombine -S | not grep {shufflevector.\*i32 8} |
Dan Gohman | ac22cfc | 2008-08-06 18:17:32 +0000 | [diff] [blame] | 2 | |
| 3 | target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" |
| 4 | target triple = "i386-apple-darwin9" |
| 5 | %struct.ActiveTextureTargets = type { i64, i64, i64, i64, i64, i64 } |
| 6 | %struct.AlphaTest = type { float, i16, i8, i8 } |
| 7 | %struct.ArrayRange = type { i8, i8, i8, i8 } |
| 8 | %struct.BlendMode = type { i16, i16, i16, i16, %struct.IColor4, i16, i16, i8, i8, i8, i8 } |
| 9 | %struct.ClearColor = type { double, %struct.IColor4, %struct.IColor4, float, i32 } |
| 10 | %struct.ClipPlane = type { i32, [6 x %struct.IColor4] } |
| 11 | %struct.ColorBuffer = type { i16, i8, i8, [8 x i16], [0 x i32] } |
| 12 | %struct.ColorMatrix = type { [16 x float]*, %struct.ImagingColorScale } |
| 13 | %struct.Convolution = type { %struct.IColor4, %struct.ImagingColorScale, i16, i16, [0 x i32], float*, i32, i32 } |
| 14 | %struct.DepthTest = type { i16, i16, i8, i8, i8, i8, double, double } |
| 15 | %struct.FixedFunction = type { %struct.PPStreamToken* } |
| 16 | %struct.FogMode = type { %struct.IColor4, float, float, float, float, float, i16, i16, i16, i8, i8 } |
| 17 | %struct.HintMode = type { i16, i16, i16, i16, i16, i16, i16, i16, i16, i16 } |
| 18 | %struct.Histogram = type { %struct.ProgramLimits*, i32, i16, i8, i8 } |
| 19 | %struct.ImagingColorScale = type { %struct.TCoord2, %struct.TCoord2, %struct.TCoord2, %struct.TCoord2 } |
| 20 | %struct.ImagingSubset = type { %struct.Convolution, %struct.Convolution, %struct.Convolution, %struct.ColorMatrix, %struct.Minmax, %struct.Histogram, %struct.ImagingColorScale, %struct.ImagingColorScale, %struct.ImagingColorScale, %struct.ImagingColorScale, i32, [0 x i32] } |
| 21 | %struct.Light = type { %struct.IColor4, %struct.IColor4, %struct.IColor4, %struct.IColor4, %struct.PointLineLimits, float, float, float, float, float, %struct.PointLineLimits, float, %struct.PointLineLimits, float, %struct.PointLineLimits, float, float, float, float, float } |
| 22 | %struct.LightModel = type { %struct.IColor4, [8 x %struct.Light], [2 x %struct.Material], i32, i16, i16, i16, i8, i8, i8, i8, i8, i8 } |
| 23 | %struct.LightProduct = type { %struct.IColor4, %struct.IColor4, %struct.IColor4 } |
| 24 | %struct.LineMode = type { float, i32, i16, i16, i8, i8, i8, i8 } |
| 25 | %struct.LogicOp = type { i16, i8, i8 } |
| 26 | %struct.MaskMode = type { i32, [3 x i32], i8, i8, i8, i8, i8, i8, i8, i8 } |
| 27 | %struct.Material = type { %struct.IColor4, %struct.IColor4, %struct.IColor4, %struct.IColor4, float, float, float, float, [8 x %struct.LightProduct], %struct.IColor4, [8 x i32] } |
| 28 | %struct.Minmax = type { %struct.MinmaxTable*, i16, i8, i8, [0 x i32] } |
| 29 | %struct.MinmaxTable = type { %struct.IColor4, %struct.IColor4 } |
| 30 | %struct.Mipmaplevel = type { [4 x i32], [4 x i32], [4 x float], [4 x i32], i32, i32, float*, i8*, i16, i16, i16, i16, [2 x float] } |
| 31 | %struct.Multisample = type { float, i8, i8, i8, i8, i8, i8, i8, i8 } |
| 32 | %struct.PipelineProgramState = type { i8, i8, i8, i8, [0 x i32], %struct.IColor4* } |
| 33 | %struct.PixelMap = type { i32*, float*, float*, float*, float*, float*, float*, float*, float*, i32*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32 } |
| 34 | %struct.PixelMode = type { float, float, %struct.PixelStore, %struct.PixelTransfer, %struct.PixelMap, %struct.ImagingSubset, i32, i32 } |
| 35 | %struct.PixelPack = type { i32, i32, i32, i32, i32, i32, i32, i32, i8, i8, i8, i8 } |
| 36 | %struct.PixelStore = type { %struct.PixelPack, %struct.PixelPack } |
| 37 | %struct.PixelTransfer = type { float, float, float, float, float, float, float, float, float, float, i32, i32, float, float, float, float, float, float, float, float, float, float, float, float } |
| 38 | %struct.PluginBufferData = type { i32 } |
| 39 | %struct.PointLineLimits = type { float, float, float } |
| 40 | %struct.PointMode = type { float, float, float, float, %struct.PointLineLimits, float, i8, i8, i8, i8, i16, i16, i32, i16, i16 } |
| 41 | %struct.PolygonMode = type { [128 x i8], float, float, i16, i16, i16, i16, i8, i8, i8, i8, i8, i8, i8, i8 } |
| 42 | %struct.ProgramLimits = type { i32, i32, i32, i32 } |
| 43 | %struct.RegisterCombiners = type { i8, i8, i8, i8, i32, [2 x %struct.IColor4], [8 x %struct.RegisterCombinersPerStageState], %struct.RegisterCombinersFinalStageState } |
| 44 | %struct.RegisterCombinersFinalStageState = type { i8, i8, i8, i8, [7 x %struct.RegisterCombinersPerVariableState] } |
| 45 | %struct.RegisterCombinersPerPortionState = type { [4 x %struct.RegisterCombinersPerVariableState], i8, i8, i8, i8, i16, i16, i16, i16, i16, i16 } |
| 46 | %struct.RegisterCombinersPerStageState = type { [2 x %struct.RegisterCombinersPerPortionState], [2 x %struct.IColor4] } |
| 47 | %struct.RegisterCombinersPerVariableState = type { i16, i16, i16, i16 } |
| 48 | %struct.SWRSurfaceRec = type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8*, i8*, i8*, [4 x i8*], i32 } |
| 49 | %struct.ScissorTest = type { %struct.ProgramLimits, i8, i8, i8, i8 } |
| 50 | %struct.State = type <{ i16, i16, i16, i16, i32, i32, [256 x %struct.IColor4], [128 x %struct.IColor4], %struct.Viewport, %struct.Transform, %struct.LightModel, %struct.ActiveTextureTargets, %struct.AlphaTest, %struct.BlendMode, %struct.ClearColor, %struct.ColorBuffer, %struct.DepthTest, %struct.ArrayRange, %struct.FogMode, %struct.HintMode, %struct.LineMode, %struct.LogicOp, %struct.MaskMode, %struct.PixelMode, %struct.PointMode, %struct.PolygonMode, %struct.ScissorTest, i32, %struct.StencilTest, [8 x %struct.TextureMode], [16 x %struct.TextureImageMode], %struct.ArrayRange, [8 x %struct.TextureCoordGen], %struct.ClipPlane, %struct.Multisample, %struct.RegisterCombiners, %struct.ArrayRange, %struct.ArrayRange, [3 x %struct.PipelineProgramState], %struct.ArrayRange, %struct.TransformFeedback, i32*, %struct.FixedFunction, [3 x i32], [3 x i32] }> |
| 51 | %struct.StencilTest = type { [3 x { i32, i32, i16, i16, i16, i16 }], i32, [4 x i8] } |
| 52 | %struct.TextureCoordGen = type { { i16, i16, %struct.IColor4, %struct.IColor4 }, { i16, i16, %struct.IColor4, %struct.IColor4 }, { i16, i16, %struct.IColor4, %struct.IColor4 }, { i16, i16, %struct.IColor4, %struct.IColor4 }, i8, i8, i8, i8 } |
| 53 | %struct.TextureGeomState = type { i16, i16, i16, i16, i16, i8, i8, i8, i8, i16, i16, i16, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, [6 x i16], [6 x i16] } |
| 54 | %struct.TextureImageMode = type { float } |
| 55 | %struct.TextureLevel = type { i32, i32, i16, i16, i16, i8, i8, i16, i16, i16, i16, i8* } |
| 56 | %struct.TextureMode = type { %struct.IColor4, i32, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, i16, float, float, i16, i16, i16, i16, i16, i16, [4 x i16], i8, i8, i8, i8, [3 x float], [4 x float], float, float } |
| 57 | %struct.TextureParamState = type { i16, i16, i16, i16, i16, i16, %struct.IColor4, float, float, float, float, i16, i16, i16, i16, float, i16, i8, i8, i32, i8* } |
| 58 | %struct.TextureRec = type { [4 x float], %struct.TextureState*, %struct.Mipmaplevel*, %struct.Mipmaplevel*, float, float, float, float, i8, i8, i8, i8, i16, i16, i16, i16, i32, float, [2 x %struct.PPStreamToken] } |
| 59 | %struct.TextureState = type { i16, i8, i8, i16, i16, float, i32, %struct.SWRSurfaceRec*, %struct.TextureParamState, %struct.TextureGeomState, [0 x i32], i8*, i32, %struct.TextureLevel, [1 x [15 x %struct.TextureLevel]] } |
| 60 | %struct.Transform = type <{ [24 x [16 x float]], [24 x [16 x float]], [16 x float], float, float, float, float, float, i8, i8, i8, i8, i32, i32, i32, i16, i16, i8, i8, i8, i8, i32 }> |
| 61 | %struct.TransformFeedback = type { i8, i8, i8, i8, [0 x i32], [16 x i32], [16 x i32] } |
| 62 | %struct.Viewport = type { float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, float, double, double, i32, i32, i32, i32, float, float, float, float } |
| 63 | %struct.IColor4 = type { float, float, float, float } |
| 64 | %struct.TCoord2 = type { float, float } |
| 65 | %struct.VMGPStack = type { [6 x <4 x float>*], <4 x float>*, i32, i32, <4 x float>*, <4 x float>**, i32, i32, i32, i32, i32, i32 } |
| 66 | %struct.VMTextures = type { [16 x %struct.TextureRec*] } |
| 67 | %struct.PPStreamToken = type { { i16, i16, i32 } } |
| 68 | %struct._VMConstants = type { <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, <4 x float>, float, float, float, float, float, float, float, float, float, float, float, float, [256 x float], [528 x i8], { void (i8*, i8*, i32, i8*)*, float (float)*, float (float)*, float (float)*, i32 (float)* } } |
| 69 | |
| 70 | define i32 @foo(%struct.State* %dst, <4 x float>* %prgrm, <4 x float>** %buffs, %struct._VMConstants* %cnstn, %struct.PPStreamToken* %pstrm, %struct.PluginBufferData* %gpctx, %struct.VMTextures* %txtrs, %struct.VMGPStack* %gpstk) nounwind { |
| 71 | bb266.i: |
| 72 | getelementptr <4 x float>* null, i32 11 ; <<4 x float>*>:0 [#uses=1] |
| 73 | load <4 x float>* %0, align 16 ; <<4 x float>>:1 [#uses=1] |
| 74 | shufflevector <4 x float> %1, <4 x float> undef, <4 x i32> < i32 0, i32 1, i32 1, i32 1 > ; <<4 x float>>:2 [#uses=1] |
| 75 | shufflevector <4 x float> %2, <4 x float> undef, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>>:3 [#uses=1] |
| 76 | shufflevector <4 x float> undef, <4 x float> undef, <4 x i32> < i32 0, i32 4, i32 1, i32 5 > ; <<4 x float>>:4 [#uses=1] |
| 77 | shufflevector <4 x float> %4, <4 x float> %3, <4 x i32> < i32 6, i32 7, i32 2, i32 3 > ; <<4 x float>>:5 [#uses=1] |
Dan Gohman | a5b9645 | 2009-06-04 22:49:04 +0000 | [diff] [blame] | 78 | fmul <4 x float> %5, zeroinitializer ; <<4 x float>>:6 [#uses=2] |
| 79 | fmul <4 x float> %6, %6 ; <<4 x float>>:7 [#uses=1] |
| 80 | fadd <4 x float> zeroinitializer, %7 ; <<4 x float>>:8 [#uses=1] |
Dan Gohman | ac22cfc | 2008-08-06 18:17:32 +0000 | [diff] [blame] | 81 | call <4 x float> @llvm.x86.sse.max.ps( <4 x float> zeroinitializer, <4 x float> %8 ) nounwind readnone ; <<4 x float>>:9 [#uses=1] |
| 82 | %phitmp40 = bitcast <4 x float> %9 to <4 x i32> ; <<4 x i32>> [#uses=1] |
| 83 | %tmp4109.i = and <4 x i32> %phitmp40, < i32 8388607, i32 8388607, i32 8388607, i32 8388607 > ; <<4 x i32>> [#uses=1] |
| 84 | %tmp4116.i = or <4 x i32> %tmp4109.i, < i32 1065353216, i32 1065353216, i32 1065353216, i32 1065353216 > ; <<4 x i32>> [#uses=1] |
| 85 | %tmp4117.i = bitcast <4 x i32> %tmp4116.i to <4 x float> ; <<4 x float>> [#uses=1] |
Dan Gohman | a5b9645 | 2009-06-04 22:49:04 +0000 | [diff] [blame] | 86 | fadd <4 x float> %tmp4117.i, zeroinitializer ; <<4 x float>>:10 [#uses=1] |
| 87 | fmul <4 x float> %10, < float 5.000000e-01, float 5.000000e-01, float 5.000000e-01, float 5.000000e-01 > ; <<4 x float>>:11 [#uses=1] |
Dan Gohman | ac22cfc | 2008-08-06 18:17:32 +0000 | [diff] [blame] | 88 | call <4 x float> @llvm.x86.sse.max.ps( <4 x float> %11, <4 x float> zeroinitializer ) nounwind readnone ; <<4 x float>>:12 [#uses=1] |
| 89 | call <4 x float> @llvm.x86.sse.min.ps( <4 x float> %12, <4 x float> zeroinitializer ) nounwind readnone ; <<4 x float>>:13 [#uses=1] |
| 90 | %tmp4170.i = call <4 x float> @llvm.x86.sse.cmp.ps( <4 x float> %13, <4 x float> zeroinitializer, i8 2 ) nounwind ; <<4 x float>> [#uses=1] |
| 91 | bitcast <4 x float> %tmp4170.i to <16 x i8> ; <<16 x i8>>:14 [#uses=1] |
| 92 | call i32 @llvm.x86.sse2.pmovmskb.128( <16 x i8> %14 ) nounwind readnone ; <i32>:15 [#uses=1] |
| 93 | icmp eq i32 %15, 0 ; <i1>:16 [#uses=1] |
| 94 | br i1 %16, label %bb5574.i, label %bb4521.i |
| 95 | |
| 96 | bb4521.i: ; preds = %bb266.i |
| 97 | unreachable |
| 98 | |
| 99 | bb5574.i: ; preds = %bb266.i |
| 100 | unreachable |
| 101 | } |
| 102 | |
| 103 | declare <4 x float> @llvm.x86.sse.cmp.ps(<4 x float>, <4 x float>, i8) nounwind readnone |
| 104 | |
| 105 | declare i32 @llvm.x86.sse2.pmovmskb.128(<16 x i8>) nounwind readnone |
| 106 | |
| 107 | declare <4 x float> @llvm.x86.sse.max.ps(<4 x float>, <4 x float>) nounwind readnone |
| 108 | |
| 109 | declare <4 x float> @llvm.x86.sse.min.ps(<4 x float>, <4 x float>) nounwind readnone |