diff options
| -rw-r--r-- | llvm/test/Transforms/SLPVectorizer/RISCV/reordered-buildvector-scalars.ll | 84 |
1 files changed, 84 insertions, 0 deletions
diff --git a/llvm/test/Transforms/SLPVectorizer/RISCV/reordered-buildvector-scalars.ll b/llvm/test/Transforms/SLPVectorizer/RISCV/reordered-buildvector-scalars.ll index f7e629f..d4e3238 100644 --- a/llvm/test/Transforms/SLPVectorizer/RISCV/reordered-buildvector-scalars.ll +++ b/llvm/test/Transforms/SLPVectorizer/RISCV/reordered-buildvector-scalars.ll @@ -1,5 +1,6 @@ ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5 ; RUN: opt -S --passes=slp-vectorizer -mtriple=riscv64-unknown-linux-gnu -mcpu=sifive-x280 < %s | FileCheck %s +; RUN: opt -S --passes=slp-vectorizer -mtriple=riscv64-unknown-linux-gnu -mcpu=sifive-x280 < %s -slp-threshold=-3 | FileCheck %s --check-prefix=THRESH %struct.ImageParameters = type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, ptr, ptr, i32, ptr, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [9 x [16 x [16 x i16]]], [5 x [16 x [16 x i16]]], [9 x [8 x [8 x i16]]], [2 x [4 x [16 x [16 x i16]]]], [16 x [16 x i16]], [16 x [16 x i32]], ptr, ptr, ptr, ptr, ptr, [1200 x %struct.syntaxelement], ptr, ptr, i32, i32, i32, i32, [4 x [4 x i32]], i32, i32, i32, i32, i32, double, i32, i32, i32, i32, ptr, ptr, ptr, ptr, [15 x i16], i32, i32, i32, i32, i32, i32, i32, i32, [6 x [15 x i32]], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [1 x i32], i32, i32, [2 x i32], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, ptr, i32, i32, i32, i32, i32, double, i32, i32, i32, i32, i32, i32, i32, ptr, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [2 x i32], i32, i32, i32 } %struct.syntaxelement = type { i32, i32, i32, i32, i32, i32, i32, i32, ptr, ptr } @@ -94,6 +95,89 @@ define fastcc i32 @test(i32 %0, i32 %add111.i.i, <4 x i32> %PredPel.i.sroa.86.72 ; CHECK-NEXT: store i16 [[CONV2206_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8218), align 2 ; CHECK-NEXT: ret i32 0 ; +; THRESH-LABEL: define fastcc i32 @test( +; THRESH-SAME: i32 [[TMP0:%.*]], i32 [[ADD111_I_I:%.*]], <4 x i32> [[PREDPEL_I_SROA_86_72_VEC_EXTRACT:%.*]], <4 x i32> [[TMP1:%.*]]) #[[ATTR0:[0-9]+]] { +; THRESH-NEXT: [[ENTRY:.*:]] +; THRESH-NEXT: [[LOOPARRAY_SROA_24_0_I_I3:%.*]] = ashr i32 [[TMP0]], 1 +; THRESH-NEXT: [[SHR143_5_I_I9:%.*]] = ashr i32 [[TMP0]], 1 +; THRESH-NEXT: [[ADD1392_I:%.*]] = add i32 [[TMP0]], 1 +; THRESH-NEXT: [[MUL1445_I:%.*]] = shl i32 [[TMP0]], 1 +; THRESH-NEXT: [[ADD2136_I:%.*]] = or i32 [[LOOPARRAY_SROA_24_0_I_I3]], [[TMP0]] +; THRESH-NEXT: [[SHR2137_I:%.*]] = lshr i32 [[ADD2136_I]], 1 +; THRESH-NEXT: [[CONV2138_I:%.*]] = trunc i32 [[SHR2137_I]] to i16 +; THRESH-NEXT: [[ADD2174_I:%.*]] = add i32 [[MUL1445_I]], 2 +; THRESH-NEXT: [[SHR2175_I:%.*]] = lshr i32 [[ADD2174_I]], 2 +; THRESH-NEXT: [[CONV2176_I:%.*]] = trunc i32 [[SHR2175_I]] to i16 +; THRESH-NEXT: [[ADD2190_I:%.*]] = or i32 [[ADD1392_I]], 1 +; THRESH-NEXT: [[ADD2191_I:%.*]] = add i32 [[ADD2190_I]], [[TMP0]] +; THRESH-NEXT: [[CONV2193_I:%.*]] = trunc i32 [[ADD2191_I]] to i16 +; THRESH-NEXT: [[ADD2203_I:%.*]] = or i32 [[TMP0]], 1 +; THRESH-NEXT: [[ADD2204_I:%.*]] = add i32 [[ADD2203_I]], [[TMP0]] +; THRESH-NEXT: [[CONV2206_I:%.*]] = trunc i32 [[ADD2204_I]] to i16 +; THRESH-NEXT: [[ADD2235_I16:%.*]] = or i32 [[TMP0]], 1 +; THRESH-NEXT: [[ADD2236_I:%.*]] = add i32 [[ADD2235_I16]], 1 +; THRESH-NEXT: [[SHR2237_I:%.*]] = lshr i32 [[ADD2236_I]], 1 +; THRESH-NEXT: [[CONV2238_I:%.*]] = trunc i32 [[SHR2237_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2238_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8196), align 4 +; THRESH-NEXT: store i16 [[CONV2238_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8176), align 8 +; THRESH-NEXT: [[ADD2258_I:%.*]] = or i32 [[ADD111_I_I]], [[TMP0]] +; THRESH-NEXT: [[SHR2259_I:%.*]] = lshr i32 [[ADD2258_I]], 1 +; THRESH-NEXT: [[CONV2260_I:%.*]] = trunc i32 [[SHR2259_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2260_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8212), align 4 +; THRESH-NEXT: store i16 [[CONV2260_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8192), align 8 +; THRESH-NEXT: store i16 [[CONV2260_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8172), align 4 +; THRESH-NEXT: [[ADD2302_I:%.*]] = add i32 [[TMP0]], 1 +; THRESH-NEXT: [[SHR2303_I:%.*]] = lshr i32 [[ADD2302_I]], 1 +; THRESH-NEXT: [[CONV2304_I:%.*]] = trunc i32 [[SHR2303_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2304_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8224), align 8 +; THRESH-NEXT: store i16 [[CONV2304_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8204), align 4 +; THRESH-NEXT: store i16 [[CONV2304_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8184), align 8 +; THRESH-NEXT: [[ADD2323_I:%.*]] = add i32 [[TMP0]], 1 +; THRESH-NEXT: [[ADD2324_I:%.*]] = or i32 [[ADD2323_I]], [[TMP0]] +; THRESH-NEXT: [[SHR2325_I:%.*]] = lshr i32 [[ADD2324_I]], 1 +; THRESH-NEXT: [[CONV2326_I:%.*]] = trunc i32 [[SHR2325_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2326_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8220), align 4 +; THRESH-NEXT: store i16 [[CONV2326_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8200), align 8 +; THRESH-NEXT: [[ADD2342_I:%.*]] = add i32 [[SHR143_5_I_I9]], 1 +; THRESH-NEXT: [[SHR2343_I:%.*]] = lshr i32 [[ADD2342_I]], 1 +; THRESH-NEXT: [[CONV2344_I:%.*]] = trunc i32 [[SHR2343_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2344_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8216), align 8 +; THRESH-NEXT: [[ADD2355_I:%.*]] = or i32 [[SHR143_5_I_I9]], 1 +; THRESH-NEXT: [[ADD2356_I:%.*]] = add i32 [[ADD2355_I]], [[TMP0]] +; THRESH-NEXT: [[CONV2358_I:%.*]] = trunc i32 [[ADD2356_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2358_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8232), align 8 +; THRESH-NEXT: [[TMP2:%.*]] = shufflevector <4 x i32> [[TMP1]], <4 x i32> poison, <2 x i32> <i32 poison, i32 0> +; THRESH-NEXT: [[TMP3:%.*]] = insertelement <2 x i32> [[TMP2]], i32 [[LOOPARRAY_SROA_24_0_I_I3]], i32 0 +; THRESH-NEXT: [[TMP4:%.*]] = add <2 x i32> [[TMP3]], splat (i32 1) +; THRESH-NEXT: [[TMP5:%.*]] = lshr <2 x i32> [[TMP4]], splat (i32 1) +; THRESH-NEXT: [[TMP6:%.*]] = trunc <2 x i32> [[TMP5]] to <2 x i16> +; THRESH-NEXT: store <2 x i16> [[TMP6]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8180), align 4 +; THRESH-NEXT: [[ADD2393_I:%.*]] = or i32 [[LOOPARRAY_SROA_24_0_I_I3]], 1 +; THRESH-NEXT: [[ADD2394_I:%.*]] = add i32 [[ADD2393_I]], [[TMP0]] +; THRESH-NEXT: [[CONV2396_I:%.*]] = trunc i32 [[ADD2394_I]] to i16 +; THRESH-NEXT: store i16 [[CONV2396_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8198), align 2 +; THRESH-NEXT: store i16 [[CONV2396_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8178), align 2 +; THRESH-NEXT: store i16 [[CONV2138_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8214), align 2 +; THRESH-NEXT: store i16 [[CONV2138_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8194), align 2 +; THRESH-NEXT: store i16 [[CONV2138_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8174), align 2 +; THRESH-NEXT: [[TMP7:%.*]] = shufflevector <4 x i32> [[PREDPEL_I_SROA_86_72_VEC_EXTRACT]], <4 x i32> poison, <2 x i32> <i32 poison, i32 0> +; THRESH-NEXT: [[TMP8:%.*]] = insertelement <2 x i32> [[TMP7]], i32 [[ADD111_I_I]], i32 0 +; THRESH-NEXT: [[TMP9:%.*]] = add <2 x i32> [[TMP8]], splat (i32 1) +; THRESH-NEXT: [[TMP10:%.*]] = lshr <2 x i32> [[TMP9]], splat (i32 1) +; THRESH-NEXT: [[TMP11:%.*]] = trunc <2 x i32> [[TMP10]] to <2 x i16> +; THRESH-NEXT: [[TMP12:%.*]] = extractelement <2 x i16> [[TMP11]], i32 1 +; THRESH-NEXT: store <2 x i16> [[TMP11]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8228), align 4 +; THRESH-NEXT: store <2 x i16> [[TMP11]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8208), align 8 +; THRESH-NEXT: store <2 x i16> [[TMP11]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8188), align 4 +; THRESH-NEXT: store i16 [[TMP12]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8170), align 2 +; THRESH-NEXT: store i16 [[CONV2176_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8226), align 2 +; THRESH-NEXT: store i16 [[CONV2176_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8206), align 2 +; THRESH-NEXT: store i16 [[CONV2176_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8186), align 2 +; THRESH-NEXT: store i16 [[CONV2193_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8222), align 2 +; THRESH-NEXT: store i16 [[CONV2193_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8202), align 2 +; THRESH-NEXT: store i16 [[CONV2206_I]], ptr getelementptr inbounds nuw (i8, ptr @images, i64 8218), align 2 +; THRESH-NEXT: ret i32 0 +; entry: %LoopArray.sroa.24.0.i.i3 = ashr i32 %0, 1 %shr143.5.i.i9 = ashr i32 %0, 1 |