| ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py |
| ; RUN: opt < %s -mtriple=x86_64-unknown -basic-aa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE --check-prefix=SSE2 |
| ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7 -basic-aa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE --check-prefix=SSE42 |
| ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -basic-aa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX1 |
| ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -basic-aa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX2 |
| |
| target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" |
| |
| @src64 = common global [4 x i64] zeroinitializer, align 32 |
| @dst64 = common global [4 x i64] zeroinitializer, align 32 |
| @src32 = common global [8 x i32] zeroinitializer, align 32 |
| @dst32 = common global [8 x i32] zeroinitializer, align 32 |
| @src16 = common global [16 x i16] zeroinitializer, align 32 |
| @dst16 = common global [16 x i16] zeroinitializer, align 32 |
| @src8 = common global [32 x i8] zeroinitializer, align 32 |
| @dst8 = common global [32 x i8] zeroinitializer, align 32 |
| |
| declare i64 @llvm.cttz.i64(i64, i1) |
| declare i32 @llvm.cttz.i32(i32, i1) |
| declare i16 @llvm.cttz.i16(i16, i1) |
| declare i8 @llvm.cttz.i8(i8, i1) |
| |
| ; |
| ; CTTZ |
| ; |
| |
| define void @cttz_2i64() #0 { |
| ; CHECK-LABEL: @cttz_2i64( |
| ; CHECK-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8 |
| ; CHECK-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8 |
| ; CHECK-NEXT: [[CTTZ0:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD0]], i1 false) |
| ; CHECK-NEXT: [[CTTZ1:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD1]], i1 false) |
| ; CHECK-NEXT: store i64 [[CTTZ0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8 |
| ; CHECK-NEXT: store i64 [[CTTZ1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8 |
| ; CHECK-NEXT: ret void |
| ; |
| %ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8 |
| %ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8 |
| %cttz0 = call i64 @llvm.cttz.i64(i64 %ld0, i1 0) |
| %cttz1 = call i64 @llvm.cttz.i64(i64 %ld1, i1 0) |
| store i64 %cttz0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8 |
| store i64 %cttz1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8 |
| ret void |
| } |
| |
| define void @cttz_4i64() #0 { |
| ; SSE-LABEL: @cttz_4i64( |
| ; SSE-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4 |
| ; SSE-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4 |
| ; SSE-NEXT: [[LD2:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4 |
| ; SSE-NEXT: [[LD3:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4 |
| ; SSE-NEXT: [[CTTZ0:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD0]], i1 false) |
| ; SSE-NEXT: [[CTTZ1:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD1]], i1 false) |
| ; SSE-NEXT: [[CTTZ2:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD2]], i1 false) |
| ; SSE-NEXT: [[CTTZ3:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD3]], i1 false) |
| ; SSE-NEXT: store i64 [[CTTZ0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4 |
| ; SSE-NEXT: store i64 [[CTTZ1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4 |
| ; SSE-NEXT: store i64 [[CTTZ2]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4 |
| ; SSE-NEXT: store i64 [[CTTZ3]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4 |
| ; SSE-NEXT: ret void |
| ; |
| ; AVX1-LABEL: @cttz_4i64( |
| ; AVX1-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4 |
| ; AVX1-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4 |
| ; AVX1-NEXT: [[LD2:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4 |
| ; AVX1-NEXT: [[LD3:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4 |
| ; AVX1-NEXT: [[CTTZ0:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD0]], i1 false) |
| ; AVX1-NEXT: [[CTTZ1:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD1]], i1 false) |
| ; AVX1-NEXT: [[CTTZ2:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD2]], i1 false) |
| ; AVX1-NEXT: [[CTTZ3:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD3]], i1 false) |
| ; AVX1-NEXT: store i64 [[CTTZ0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4 |
| ; AVX1-NEXT: store i64 [[CTTZ1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4 |
| ; AVX1-NEXT: store i64 [[CTTZ2]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4 |
| ; AVX1-NEXT: store i64 [[CTTZ3]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4 |
| ; AVX1-NEXT: ret void |
| ; |
| ; AVX2-LABEL: @cttz_4i64( |
| ; AVX2-NEXT: [[TMP1:%.*]] = load <4 x i64>, <4 x i64>* bitcast ([4 x i64]* @src64 to <4 x i64>*), align 4 |
| ; AVX2-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.cttz.v4i64(<4 x i64> [[TMP1]], i1 false) |
| ; AVX2-NEXT: store <4 x i64> [[TMP2]], <4 x i64>* bitcast ([4 x i64]* @dst64 to <4 x i64>*), align 4 |
| ; AVX2-NEXT: ret void |
| ; |
| %ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4 |
| %ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4 |
| %ld2 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4 |
| %ld3 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4 |
| %cttz0 = call i64 @llvm.cttz.i64(i64 %ld0, i1 0) |
| %cttz1 = call i64 @llvm.cttz.i64(i64 %ld1, i1 0) |
| %cttz2 = call i64 @llvm.cttz.i64(i64 %ld2, i1 0) |
| %cttz3 = call i64 @llvm.cttz.i64(i64 %ld3, i1 0) |
| store i64 %cttz0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4 |
| store i64 %cttz1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4 |
| store i64 %cttz2, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4 |
| store i64 %cttz3, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4 |
| ret void |
| } |
| |
| define void @cttz_4i32() #0 { |
| ; SSE2-LABEL: @cttz_4i32( |
| ; SSE2-NEXT: [[LD0:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 4 |
| ; SSE2-NEXT: [[LD1:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 4 |
| ; SSE2-NEXT: [[LD2:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 4 |
| ; SSE2-NEXT: [[LD3:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 4 |
| ; SSE2-NEXT: [[CTTZ0:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD0]], i1 false) |
| ; SSE2-NEXT: [[CTTZ1:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD1]], i1 false) |
| ; SSE2-NEXT: [[CTTZ2:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD2]], i1 false) |
| ; SSE2-NEXT: [[CTTZ3:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD3]], i1 false) |
| ; SSE2-NEXT: store i32 [[CTTZ0]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 4 |
| ; SSE2-NEXT: store i32 [[CTTZ1]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 4 |
| ; SSE2-NEXT: store i32 [[CTTZ2]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 4 |
| ; SSE2-NEXT: store i32 [[CTTZ3]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 4 |
| ; SSE2-NEXT: ret void |
| ; |
| ; SSE42-LABEL: @cttz_4i32( |
| ; SSE42-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 4 |
| ; SSE42-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP1]], i1 false) |
| ; SSE42-NEXT: store <4 x i32> [[TMP2]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 4 |
| ; SSE42-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_4i32( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 4 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP1]], i1 false) |
| ; AVX-NEXT: store <4 x i32> [[TMP2]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 4 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 4 |
| %ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 4 |
| %ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 4 |
| %ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 4 |
| %cttz0 = call i32 @llvm.cttz.i32(i32 %ld0, i1 0) |
| %cttz1 = call i32 @llvm.cttz.i32(i32 %ld1, i1 0) |
| %cttz2 = call i32 @llvm.cttz.i32(i32 %ld2, i1 0) |
| %cttz3 = call i32 @llvm.cttz.i32(i32 %ld3, i1 0) |
| store i32 %cttz0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 4 |
| store i32 %cttz1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 4 |
| store i32 %cttz2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 4 |
| store i32 %cttz3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 4 |
| ret void |
| } |
| |
| define void @cttz_8i32() #0 { |
| ; SSE2-LABEL: @cttz_8i32( |
| ; SSE2-NEXT: [[LD0:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 2 |
| ; SSE2-NEXT: [[LD1:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 2 |
| ; SSE2-NEXT: [[LD2:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 2 |
| ; SSE2-NEXT: [[LD3:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 2 |
| ; SSE2-NEXT: [[LD4:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4), align 2 |
| ; SSE2-NEXT: [[LD5:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 5), align 2 |
| ; SSE2-NEXT: [[LD6:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 6), align 2 |
| ; SSE2-NEXT: [[LD7:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 7), align 2 |
| ; SSE2-NEXT: [[CTTZ0:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD0]], i1 false) |
| ; SSE2-NEXT: [[CTTZ1:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD1]], i1 false) |
| ; SSE2-NEXT: [[CTTZ2:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD2]], i1 false) |
| ; SSE2-NEXT: [[CTTZ3:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD3]], i1 false) |
| ; SSE2-NEXT: [[CTTZ4:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD4]], i1 false) |
| ; SSE2-NEXT: [[CTTZ5:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD5]], i1 false) |
| ; SSE2-NEXT: [[CTTZ6:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD6]], i1 false) |
| ; SSE2-NEXT: [[CTTZ7:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD7]], i1 false) |
| ; SSE2-NEXT: store i32 [[CTTZ0]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ1]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ2]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ3]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ4]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ5]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 5), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ6]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 6), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ7]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 7), align 2 |
| ; SSE2-NEXT: ret void |
| ; |
| ; SSE42-LABEL: @cttz_8i32( |
| ; SSE42-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 2 |
| ; SSE42-NEXT: [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4) to <4 x i32>*), align 2 |
| ; SSE42-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP1]], i1 false) |
| ; SSE42-NEXT: [[TMP4:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP2]], i1 false) |
| ; SSE42-NEXT: store <4 x i32> [[TMP3]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 2 |
| ; SSE42-NEXT: store <4 x i32> [[TMP4]], <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4) to <4 x i32>*), align 2 |
| ; SSE42-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_8i32( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([8 x i32]* @src32 to <8 x i32>*), align 2 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.cttz.v8i32(<8 x i32> [[TMP1]], i1 false) |
| ; AVX-NEXT: store <8 x i32> [[TMP2]], <8 x i32>* bitcast ([8 x i32]* @dst32 to <8 x i32>*), align 2 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 2 |
| %ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 2 |
| %ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 2 |
| %ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 2 |
| %ld4 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4), align 2 |
| %ld5 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 5), align 2 |
| %ld6 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 6), align 2 |
| %ld7 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 7), align 2 |
| %cttz0 = call i32 @llvm.cttz.i32(i32 %ld0, i1 0) |
| %cttz1 = call i32 @llvm.cttz.i32(i32 %ld1, i1 0) |
| %cttz2 = call i32 @llvm.cttz.i32(i32 %ld2, i1 0) |
| %cttz3 = call i32 @llvm.cttz.i32(i32 %ld3, i1 0) |
| %cttz4 = call i32 @llvm.cttz.i32(i32 %ld4, i1 0) |
| %cttz5 = call i32 @llvm.cttz.i32(i32 %ld5, i1 0) |
| %cttz6 = call i32 @llvm.cttz.i32(i32 %ld6, i1 0) |
| %cttz7 = call i32 @llvm.cttz.i32(i32 %ld7, i1 0) |
| store i32 %cttz0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 2 |
| store i32 %cttz1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 2 |
| store i32 %cttz2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 2 |
| store i32 %cttz3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 2 |
| store i32 %cttz4, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4), align 2 |
| store i32 %cttz5, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 5), align 2 |
| store i32 %cttz6, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 6), align 2 |
| store i32 %cttz7, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 7), align 2 |
| ret void |
| } |
| |
| define void @cttz_8i16() #0 { |
| ; CHECK-LABEL: @cttz_8i16( |
| ; CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2 |
| ; CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[TMP1]], i1 false) |
| ; CHECK-NEXT: store <8 x i16> [[TMP2]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2 |
| ; CHECK-NEXT: ret void |
| ; |
| %ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2 |
| %ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2 |
| %ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2 |
| %ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2 |
| %ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2 |
| %ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2 |
| %ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2 |
| %ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2 |
| %cttz0 = call i16 @llvm.cttz.i16(i16 %ld0, i1 0) |
| %cttz1 = call i16 @llvm.cttz.i16(i16 %ld1, i1 0) |
| %cttz2 = call i16 @llvm.cttz.i16(i16 %ld2, i1 0) |
| %cttz3 = call i16 @llvm.cttz.i16(i16 %ld3, i1 0) |
| %cttz4 = call i16 @llvm.cttz.i16(i16 %ld4, i1 0) |
| %cttz5 = call i16 @llvm.cttz.i16(i16 %ld5, i1 0) |
| %cttz6 = call i16 @llvm.cttz.i16(i16 %ld6, i1 0) |
| %cttz7 = call i16 @llvm.cttz.i16(i16 %ld7, i1 0) |
| store i16 %cttz0, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2 |
| store i16 %cttz1, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2 |
| store i16 %cttz2, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2 |
| store i16 %cttz3, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2 |
| store i16 %cttz4, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2 |
| store i16 %cttz5, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2 |
| store i16 %cttz6, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2 |
| store i16 %cttz7, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2 |
| ret void |
| } |
| |
| define void @cttz_16i16() #0 { |
| ; SSE-LABEL: @cttz_16i16( |
| ; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2 |
| ; SSE-NEXT: [[TMP2:%.*]] = load <8 x i16>, <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8) to <8 x i16>*), align 2 |
| ; SSE-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[TMP1]], i1 false) |
| ; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[TMP2]], i1 false) |
| ; SSE-NEXT: store <8 x i16> [[TMP3]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2 |
| ; SSE-NEXT: store <8 x i16> [[TMP4]], <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8) to <8 x i16>*), align 2 |
| ; SSE-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_16i16( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, <16 x i16>* bitcast ([16 x i16]* @src16 to <16 x i16>*), align 2 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.cttz.v16i16(<16 x i16> [[TMP1]], i1 false) |
| ; AVX-NEXT: store <16 x i16> [[TMP2]], <16 x i16>* bitcast ([16 x i16]* @dst16 to <16 x i16>*), align 2 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2 |
| %ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2 |
| %ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2 |
| %ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2 |
| %ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2 |
| %ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2 |
| %ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2 |
| %ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2 |
| %ld8 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8), align 2 |
| %ld9 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 9), align 2 |
| %ld10 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 10), align 2 |
| %ld11 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 11), align 2 |
| %ld12 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 12), align 2 |
| %ld13 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 13), align 2 |
| %ld14 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 14), align 2 |
| %ld15 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 15), align 2 |
| %cttz0 = call i16 @llvm.cttz.i16(i16 %ld0, i1 0) |
| %cttz1 = call i16 @llvm.cttz.i16(i16 %ld1, i1 0) |
| %cttz2 = call i16 @llvm.cttz.i16(i16 %ld2, i1 0) |
| %cttz3 = call i16 @llvm.cttz.i16(i16 %ld3, i1 0) |
| %cttz4 = call i16 @llvm.cttz.i16(i16 %ld4, i1 0) |
| %cttz5 = call i16 @llvm.cttz.i16(i16 %ld5, i1 0) |
| %cttz6 = call i16 @llvm.cttz.i16(i16 %ld6, i1 0) |
| %cttz7 = call i16 @llvm.cttz.i16(i16 %ld7, i1 0) |
| %cttz8 = call i16 @llvm.cttz.i16(i16 %ld8, i1 0) |
| %cttz9 = call i16 @llvm.cttz.i16(i16 %ld9, i1 0) |
| %cttz10 = call i16 @llvm.cttz.i16(i16 %ld10, i1 0) |
| %cttz11 = call i16 @llvm.cttz.i16(i16 %ld11, i1 0) |
| %cttz12 = call i16 @llvm.cttz.i16(i16 %ld12, i1 0) |
| %cttz13 = call i16 @llvm.cttz.i16(i16 %ld13, i1 0) |
| %cttz14 = call i16 @llvm.cttz.i16(i16 %ld14, i1 0) |
| %cttz15 = call i16 @llvm.cttz.i16(i16 %ld15, i1 0) |
| store i16 %cttz0 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2 |
| store i16 %cttz1 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2 |
| store i16 %cttz2 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2 |
| store i16 %cttz3 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2 |
| store i16 %cttz4 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2 |
| store i16 %cttz5 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2 |
| store i16 %cttz6 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2 |
| store i16 %cttz7 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2 |
| store i16 %cttz8 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8), align 2 |
| store i16 %cttz9 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 9), align 2 |
| store i16 %cttz10, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 10), align 2 |
| store i16 %cttz11, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 11), align 2 |
| store i16 %cttz12, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 12), align 2 |
| store i16 %cttz13, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 13), align 2 |
| store i16 %cttz14, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 14), align 2 |
| store i16 %cttz15, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 15), align 2 |
| ret void |
| } |
| |
| define void @cttz_16i8() #0 { |
| ; CHECK-LABEL: @cttz_16i8( |
| ; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, <16 x i8>* bitcast ([32 x i8]* @src8 to <16 x i8>*), align 1 |
| ; CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.cttz.v16i8(<16 x i8> [[TMP1]], i1 false) |
| ; CHECK-NEXT: store <16 x i8> [[TMP2]], <16 x i8>* bitcast ([32 x i8]* @dst8 to <16 x i8>*), align 1 |
| ; CHECK-NEXT: ret void |
| ; |
| %ld0 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 0), align 1 |
| %ld1 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 1), align 1 |
| %ld2 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 2), align 1 |
| %ld3 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 3), align 1 |
| %ld4 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 4), align 1 |
| %ld5 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 5), align 1 |
| %ld6 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 6), align 1 |
| %ld7 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 7), align 1 |
| %ld8 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 8), align 1 |
| %ld9 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 9), align 1 |
| %ld10 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 10), align 1 |
| %ld11 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 11), align 1 |
| %ld12 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 12), align 1 |
| %ld13 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 13), align 1 |
| %ld14 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 14), align 1 |
| %ld15 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 15), align 1 |
| %cttz0 = call i8 @llvm.cttz.i8(i8 %ld0, i1 0) |
| %cttz1 = call i8 @llvm.cttz.i8(i8 %ld1, i1 0) |
| %cttz2 = call i8 @llvm.cttz.i8(i8 %ld2, i1 0) |
| %cttz3 = call i8 @llvm.cttz.i8(i8 %ld3, i1 0) |
| %cttz4 = call i8 @llvm.cttz.i8(i8 %ld4, i1 0) |
| %cttz5 = call i8 @llvm.cttz.i8(i8 %ld5, i1 0) |
| %cttz6 = call i8 @llvm.cttz.i8(i8 %ld6, i1 0) |
| %cttz7 = call i8 @llvm.cttz.i8(i8 %ld7, i1 0) |
| %cttz8 = call i8 @llvm.cttz.i8(i8 %ld8, i1 0) |
| %cttz9 = call i8 @llvm.cttz.i8(i8 %ld9, i1 0) |
| %cttz10 = call i8 @llvm.cttz.i8(i8 %ld10, i1 0) |
| %cttz11 = call i8 @llvm.cttz.i8(i8 %ld11, i1 0) |
| %cttz12 = call i8 @llvm.cttz.i8(i8 %ld12, i1 0) |
| %cttz13 = call i8 @llvm.cttz.i8(i8 %ld13, i1 0) |
| %cttz14 = call i8 @llvm.cttz.i8(i8 %ld14, i1 0) |
| %cttz15 = call i8 @llvm.cttz.i8(i8 %ld15, i1 0) |
| store i8 %cttz0 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 0), align 1 |
| store i8 %cttz1 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 1), align 1 |
| store i8 %cttz2 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 2), align 1 |
| store i8 %cttz3 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 3), align 1 |
| store i8 %cttz4 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 4), align 1 |
| store i8 %cttz5 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 5), align 1 |
| store i8 %cttz6 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 6), align 1 |
| store i8 %cttz7 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 7), align 1 |
| store i8 %cttz8 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 8), align 1 |
| store i8 %cttz9 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 9), align 1 |
| store i8 %cttz10, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 10), align 1 |
| store i8 %cttz11, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 11), align 1 |
| store i8 %cttz12, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 12), align 1 |
| store i8 %cttz13, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 13), align 1 |
| store i8 %cttz14, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 14), align 1 |
| store i8 %cttz15, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 15), align 1 |
| ret void |
| } |
| |
| define void @cttz_32i8() #0 { |
| ; SSE-LABEL: @cttz_32i8( |
| ; SSE-NEXT: [[TMP1:%.*]] = load <16 x i8>, <16 x i8>* bitcast ([32 x i8]* @src8 to <16 x i8>*), align 1 |
| ; SSE-NEXT: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* bitcast (i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 16) to <16 x i8>*), align 1 |
| ; SSE-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.cttz.v16i8(<16 x i8> [[TMP1]], i1 false) |
| ; SSE-NEXT: [[TMP4:%.*]] = call <16 x i8> @llvm.cttz.v16i8(<16 x i8> [[TMP2]], i1 false) |
| ; SSE-NEXT: store <16 x i8> [[TMP3]], <16 x i8>* bitcast ([32 x i8]* @dst8 to <16 x i8>*), align 1 |
| ; SSE-NEXT: store <16 x i8> [[TMP4]], <16 x i8>* bitcast (i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 16) to <16 x i8>*), align 1 |
| ; SSE-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_32i8( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <32 x i8>, <32 x i8>* bitcast ([32 x i8]* @src8 to <32 x i8>*), align 1 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <32 x i8> @llvm.cttz.v32i8(<32 x i8> [[TMP1]], i1 false) |
| ; AVX-NEXT: store <32 x i8> [[TMP2]], <32 x i8>* bitcast ([32 x i8]* @dst8 to <32 x i8>*), align 1 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 0), align 1 |
| %ld1 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 1), align 1 |
| %ld2 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 2), align 1 |
| %ld3 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 3), align 1 |
| %ld4 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 4), align 1 |
| %ld5 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 5), align 1 |
| %ld6 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 6), align 1 |
| %ld7 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 7), align 1 |
| %ld8 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 8), align 1 |
| %ld9 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 9), align 1 |
| %ld10 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 10), align 1 |
| %ld11 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 11), align 1 |
| %ld12 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 12), align 1 |
| %ld13 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 13), align 1 |
| %ld14 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 14), align 1 |
| %ld15 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 15), align 1 |
| %ld16 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 16), align 1 |
| %ld17 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 17), align 1 |
| %ld18 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 18), align 1 |
| %ld19 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 19), align 1 |
| %ld20 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 20), align 1 |
| %ld21 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 21), align 1 |
| %ld22 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 22), align 1 |
| %ld23 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 23), align 1 |
| %ld24 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 24), align 1 |
| %ld25 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 25), align 1 |
| %ld26 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 26), align 1 |
| %ld27 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 27), align 1 |
| %ld28 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 28), align 1 |
| %ld29 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 29), align 1 |
| %ld30 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 30), align 1 |
| %ld31 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 31), align 1 |
| %cttz0 = call i8 @llvm.cttz.i8(i8 %ld0, i1 0) |
| %cttz1 = call i8 @llvm.cttz.i8(i8 %ld1, i1 0) |
| %cttz2 = call i8 @llvm.cttz.i8(i8 %ld2, i1 0) |
| %cttz3 = call i8 @llvm.cttz.i8(i8 %ld3, i1 0) |
| %cttz4 = call i8 @llvm.cttz.i8(i8 %ld4, i1 0) |
| %cttz5 = call i8 @llvm.cttz.i8(i8 %ld5, i1 0) |
| %cttz6 = call i8 @llvm.cttz.i8(i8 %ld6, i1 0) |
| %cttz7 = call i8 @llvm.cttz.i8(i8 %ld7, i1 0) |
| %cttz8 = call i8 @llvm.cttz.i8(i8 %ld8, i1 0) |
| %cttz9 = call i8 @llvm.cttz.i8(i8 %ld9, i1 0) |
| %cttz10 = call i8 @llvm.cttz.i8(i8 %ld10, i1 0) |
| %cttz11 = call i8 @llvm.cttz.i8(i8 %ld11, i1 0) |
| %cttz12 = call i8 @llvm.cttz.i8(i8 %ld12, i1 0) |
| %cttz13 = call i8 @llvm.cttz.i8(i8 %ld13, i1 0) |
| %cttz14 = call i8 @llvm.cttz.i8(i8 %ld14, i1 0) |
| %cttz15 = call i8 @llvm.cttz.i8(i8 %ld15, i1 0) |
| %cttz16 = call i8 @llvm.cttz.i8(i8 %ld16, i1 0) |
| %cttz17 = call i8 @llvm.cttz.i8(i8 %ld17, i1 0) |
| %cttz18 = call i8 @llvm.cttz.i8(i8 %ld18, i1 0) |
| %cttz19 = call i8 @llvm.cttz.i8(i8 %ld19, i1 0) |
| %cttz20 = call i8 @llvm.cttz.i8(i8 %ld20, i1 0) |
| %cttz21 = call i8 @llvm.cttz.i8(i8 %ld21, i1 0) |
| %cttz22 = call i8 @llvm.cttz.i8(i8 %ld22, i1 0) |
| %cttz23 = call i8 @llvm.cttz.i8(i8 %ld23, i1 0) |
| %cttz24 = call i8 @llvm.cttz.i8(i8 %ld24, i1 0) |
| %cttz25 = call i8 @llvm.cttz.i8(i8 %ld25, i1 0) |
| %cttz26 = call i8 @llvm.cttz.i8(i8 %ld26, i1 0) |
| %cttz27 = call i8 @llvm.cttz.i8(i8 %ld27, i1 0) |
| %cttz28 = call i8 @llvm.cttz.i8(i8 %ld28, i1 0) |
| %cttz29 = call i8 @llvm.cttz.i8(i8 %ld29, i1 0) |
| %cttz30 = call i8 @llvm.cttz.i8(i8 %ld30, i1 0) |
| %cttz31 = call i8 @llvm.cttz.i8(i8 %ld31, i1 0) |
| store i8 %cttz0 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 0), align 1 |
| store i8 %cttz1 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 1), align 1 |
| store i8 %cttz2 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 2), align 1 |
| store i8 %cttz3 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 3), align 1 |
| store i8 %cttz4 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 4), align 1 |
| store i8 %cttz5 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 5), align 1 |
| store i8 %cttz6 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 6), align 1 |
| store i8 %cttz7 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 7), align 1 |
| store i8 %cttz8 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 8), align 1 |
| store i8 %cttz9 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 9), align 1 |
| store i8 %cttz10, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 10), align 1 |
| store i8 %cttz11, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 11), align 1 |
| store i8 %cttz12, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 12), align 1 |
| store i8 %cttz13, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 13), align 1 |
| store i8 %cttz14, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 14), align 1 |
| store i8 %cttz15, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 15), align 1 |
| store i8 %cttz16, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 16), align 1 |
| store i8 %cttz17, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 17), align 1 |
| store i8 %cttz18, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 18), align 1 |
| store i8 %cttz19, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 19), align 1 |
| store i8 %cttz20, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 20), align 1 |
| store i8 %cttz21, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 21), align 1 |
| store i8 %cttz22, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 22), align 1 |
| store i8 %cttz23, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 23), align 1 |
| store i8 %cttz24, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 24), align 1 |
| store i8 %cttz25, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 25), align 1 |
| store i8 %cttz26, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 26), align 1 |
| store i8 %cttz27, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 27), align 1 |
| store i8 %cttz28, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 28), align 1 |
| store i8 %cttz29, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 29), align 1 |
| store i8 %cttz30, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 30), align 1 |
| store i8 %cttz31, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 31), align 1 |
| ret void |
| } |
| |
| ; |
| ; CTTZ_ZERO_UNDEF |
| ; |
| |
| define void @cttz_undef_2i64() #0 { |
| ; CHECK-LABEL: @cttz_undef_2i64( |
| ; CHECK-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8 |
| ; CHECK-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8 |
| ; CHECK-NEXT: [[CTTZ0:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD0]], i1 true) |
| ; CHECK-NEXT: [[CTTZ1:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD1]], i1 true) |
| ; CHECK-NEXT: store i64 [[CTTZ0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8 |
| ; CHECK-NEXT: store i64 [[CTTZ1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8 |
| ; CHECK-NEXT: ret void |
| ; |
| %ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 0), align 8 |
| %ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i32 0, i64 1), align 8 |
| %cttz0 = call i64 @llvm.cttz.i64(i64 %ld0, i1 -1) |
| %cttz1 = call i64 @llvm.cttz.i64(i64 %ld1, i1 -1) |
| store i64 %cttz0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 0), align 8 |
| store i64 %cttz1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i32 0, i64 1), align 8 |
| ret void |
| } |
| |
| define void @cttz_undef_4i64() #0 { |
| ; SSE-LABEL: @cttz_undef_4i64( |
| ; SSE-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4 |
| ; SSE-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4 |
| ; SSE-NEXT: [[LD2:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4 |
| ; SSE-NEXT: [[LD3:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4 |
| ; SSE-NEXT: [[CTTZ0:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD0]], i1 true) |
| ; SSE-NEXT: [[CTTZ1:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD1]], i1 true) |
| ; SSE-NEXT: [[CTTZ2:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD2]], i1 true) |
| ; SSE-NEXT: [[CTTZ3:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD3]], i1 true) |
| ; SSE-NEXT: store i64 [[CTTZ0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4 |
| ; SSE-NEXT: store i64 [[CTTZ1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4 |
| ; SSE-NEXT: store i64 [[CTTZ2]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4 |
| ; SSE-NEXT: store i64 [[CTTZ3]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4 |
| ; SSE-NEXT: ret void |
| ; |
| ; AVX1-LABEL: @cttz_undef_4i64( |
| ; AVX1-NEXT: [[LD0:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4 |
| ; AVX1-NEXT: [[LD1:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4 |
| ; AVX1-NEXT: [[LD2:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4 |
| ; AVX1-NEXT: [[LD3:%.*]] = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4 |
| ; AVX1-NEXT: [[CTTZ0:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD0]], i1 true) |
| ; AVX1-NEXT: [[CTTZ1:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD1]], i1 true) |
| ; AVX1-NEXT: [[CTTZ2:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD2]], i1 true) |
| ; AVX1-NEXT: [[CTTZ3:%.*]] = call i64 @llvm.cttz.i64(i64 [[LD3]], i1 true) |
| ; AVX1-NEXT: store i64 [[CTTZ0]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4 |
| ; AVX1-NEXT: store i64 [[CTTZ1]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4 |
| ; AVX1-NEXT: store i64 [[CTTZ2]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4 |
| ; AVX1-NEXT: store i64 [[CTTZ3]], i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4 |
| ; AVX1-NEXT: ret void |
| ; |
| ; AVX2-LABEL: @cttz_undef_4i64( |
| ; AVX2-NEXT: [[TMP1:%.*]] = load <4 x i64>, <4 x i64>* bitcast ([4 x i64]* @src64 to <4 x i64>*), align 4 |
| ; AVX2-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.cttz.v4i64(<4 x i64> [[TMP1]], i1 true) |
| ; AVX2-NEXT: store <4 x i64> [[TMP2]], <4 x i64>* bitcast ([4 x i64]* @dst64 to <4 x i64>*), align 4 |
| ; AVX2-NEXT: ret void |
| ; |
| %ld0 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 0), align 4 |
| %ld1 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 1), align 4 |
| %ld2 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 2), align 4 |
| %ld3 = load i64, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @src64, i64 0, i64 3), align 4 |
| %cttz0 = call i64 @llvm.cttz.i64(i64 %ld0, i1 -1) |
| %cttz1 = call i64 @llvm.cttz.i64(i64 %ld1, i1 -1) |
| %cttz2 = call i64 @llvm.cttz.i64(i64 %ld2, i1 -1) |
| %cttz3 = call i64 @llvm.cttz.i64(i64 %ld3, i1 -1) |
| store i64 %cttz0, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 0), align 4 |
| store i64 %cttz1, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 1), align 4 |
| store i64 %cttz2, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 2), align 4 |
| store i64 %cttz3, i64* getelementptr inbounds ([4 x i64], [4 x i64]* @dst64, i64 0, i64 3), align 4 |
| ret void |
| } |
| |
| define void @cttz_undef_4i32() #0 { |
| ; SSE2-LABEL: @cttz_undef_4i32( |
| ; SSE2-NEXT: [[LD0:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 4 |
| ; SSE2-NEXT: [[LD1:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 4 |
| ; SSE2-NEXT: [[LD2:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 4 |
| ; SSE2-NEXT: [[LD3:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 4 |
| ; SSE2-NEXT: [[CTTZ0:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD0]], i1 true) |
| ; SSE2-NEXT: [[CTTZ1:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD1]], i1 true) |
| ; SSE2-NEXT: [[CTTZ2:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD2]], i1 true) |
| ; SSE2-NEXT: [[CTTZ3:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD3]], i1 true) |
| ; SSE2-NEXT: store i32 [[CTTZ0]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 4 |
| ; SSE2-NEXT: store i32 [[CTTZ1]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 4 |
| ; SSE2-NEXT: store i32 [[CTTZ2]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 4 |
| ; SSE2-NEXT: store i32 [[CTTZ3]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 4 |
| ; SSE2-NEXT: ret void |
| ; |
| ; SSE42-LABEL: @cttz_undef_4i32( |
| ; SSE42-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 4 |
| ; SSE42-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP1]], i1 true) |
| ; SSE42-NEXT: store <4 x i32> [[TMP2]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 4 |
| ; SSE42-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_undef_4i32( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 4 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP1]], i1 true) |
| ; AVX-NEXT: store <4 x i32> [[TMP2]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 4 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 4 |
| %ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 4 |
| %ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 4 |
| %ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 4 |
| %cttz0 = call i32 @llvm.cttz.i32(i32 %ld0, i1 -1) |
| %cttz1 = call i32 @llvm.cttz.i32(i32 %ld1, i1 -1) |
| %cttz2 = call i32 @llvm.cttz.i32(i32 %ld2, i1 -1) |
| %cttz3 = call i32 @llvm.cttz.i32(i32 %ld3, i1 -1) |
| store i32 %cttz0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 4 |
| store i32 %cttz1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 4 |
| store i32 %cttz2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 4 |
| store i32 %cttz3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 4 |
| ret void |
| } |
| |
| define void @cttz_undef_8i32() #0 { |
| ; SSE2-LABEL: @cttz_undef_8i32( |
| ; SSE2-NEXT: [[LD0:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 2 |
| ; SSE2-NEXT: [[LD1:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 2 |
| ; SSE2-NEXT: [[LD2:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 2 |
| ; SSE2-NEXT: [[LD3:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 2 |
| ; SSE2-NEXT: [[LD4:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4), align 2 |
| ; SSE2-NEXT: [[LD5:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 5), align 2 |
| ; SSE2-NEXT: [[LD6:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 6), align 2 |
| ; SSE2-NEXT: [[LD7:%.*]] = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 7), align 2 |
| ; SSE2-NEXT: [[CTTZ0:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD0]], i1 true) |
| ; SSE2-NEXT: [[CTTZ1:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD1]], i1 true) |
| ; SSE2-NEXT: [[CTTZ2:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD2]], i1 true) |
| ; SSE2-NEXT: [[CTTZ3:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD3]], i1 true) |
| ; SSE2-NEXT: [[CTTZ4:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD4]], i1 true) |
| ; SSE2-NEXT: [[CTTZ5:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD5]], i1 true) |
| ; SSE2-NEXT: [[CTTZ6:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD6]], i1 true) |
| ; SSE2-NEXT: [[CTTZ7:%.*]] = call i32 @llvm.cttz.i32(i32 [[LD7]], i1 true) |
| ; SSE2-NEXT: store i32 [[CTTZ0]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ1]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ2]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ3]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ4]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ5]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 5), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ6]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 6), align 2 |
| ; SSE2-NEXT: store i32 [[CTTZ7]], i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 7), align 2 |
| ; SSE2-NEXT: ret void |
| ; |
| ; SSE42-LABEL: @cttz_undef_8i32( |
| ; SSE42-NEXT: [[TMP1:%.*]] = load <4 x i32>, <4 x i32>* bitcast ([8 x i32]* @src32 to <4 x i32>*), align 2 |
| ; SSE42-NEXT: [[TMP2:%.*]] = load <4 x i32>, <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4) to <4 x i32>*), align 2 |
| ; SSE42-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP1]], i1 true) |
| ; SSE42-NEXT: [[TMP4:%.*]] = call <4 x i32> @llvm.cttz.v4i32(<4 x i32> [[TMP2]], i1 true) |
| ; SSE42-NEXT: store <4 x i32> [[TMP3]], <4 x i32>* bitcast ([8 x i32]* @dst32 to <4 x i32>*), align 2 |
| ; SSE42-NEXT: store <4 x i32> [[TMP4]], <4 x i32>* bitcast (i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4) to <4 x i32>*), align 2 |
| ; SSE42-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_undef_8i32( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, <8 x i32>* bitcast ([8 x i32]* @src32 to <8 x i32>*), align 2 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.cttz.v8i32(<8 x i32> [[TMP1]], i1 true) |
| ; AVX-NEXT: store <8 x i32> [[TMP2]], <8 x i32>* bitcast ([8 x i32]* @dst32 to <8 x i32>*), align 2 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 0), align 2 |
| %ld1 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 1), align 2 |
| %ld2 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 2), align 2 |
| %ld3 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 3), align 2 |
| %ld4 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 4), align 2 |
| %ld5 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 5), align 2 |
| %ld6 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 6), align 2 |
| %ld7 = load i32, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @src32, i32 0, i64 7), align 2 |
| %cttz0 = call i32 @llvm.cttz.i32(i32 %ld0, i1 -1) |
| %cttz1 = call i32 @llvm.cttz.i32(i32 %ld1, i1 -1) |
| %cttz2 = call i32 @llvm.cttz.i32(i32 %ld2, i1 -1) |
| %cttz3 = call i32 @llvm.cttz.i32(i32 %ld3, i1 -1) |
| %cttz4 = call i32 @llvm.cttz.i32(i32 %ld4, i1 -1) |
| %cttz5 = call i32 @llvm.cttz.i32(i32 %ld5, i1 -1) |
| %cttz6 = call i32 @llvm.cttz.i32(i32 %ld6, i1 -1) |
| %cttz7 = call i32 @llvm.cttz.i32(i32 %ld7, i1 -1) |
| store i32 %cttz0, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 0), align 2 |
| store i32 %cttz1, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 1), align 2 |
| store i32 %cttz2, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 2), align 2 |
| store i32 %cttz3, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 3), align 2 |
| store i32 %cttz4, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 4), align 2 |
| store i32 %cttz5, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 5), align 2 |
| store i32 %cttz6, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 6), align 2 |
| store i32 %cttz7, i32* getelementptr inbounds ([8 x i32], [8 x i32]* @dst32, i32 0, i64 7), align 2 |
| ret void |
| } |
| |
| define void @cttz_undef_8i16() #0 { |
| ; CHECK-LABEL: @cttz_undef_8i16( |
| ; CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2 |
| ; CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[TMP1]], i1 true) |
| ; CHECK-NEXT: store <8 x i16> [[TMP2]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2 |
| ; CHECK-NEXT: ret void |
| ; |
| %ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2 |
| %ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2 |
| %ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2 |
| %ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2 |
| %ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2 |
| %ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2 |
| %ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2 |
| %ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2 |
| %cttz0 = call i16 @llvm.cttz.i16(i16 %ld0, i1 -1) |
| %cttz1 = call i16 @llvm.cttz.i16(i16 %ld1, i1 -1) |
| %cttz2 = call i16 @llvm.cttz.i16(i16 %ld2, i1 -1) |
| %cttz3 = call i16 @llvm.cttz.i16(i16 %ld3, i1 -1) |
| %cttz4 = call i16 @llvm.cttz.i16(i16 %ld4, i1 -1) |
| %cttz5 = call i16 @llvm.cttz.i16(i16 %ld5, i1 -1) |
| %cttz6 = call i16 @llvm.cttz.i16(i16 %ld6, i1 -1) |
| %cttz7 = call i16 @llvm.cttz.i16(i16 %ld7, i1 -1) |
| store i16 %cttz0, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2 |
| store i16 %cttz1, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2 |
| store i16 %cttz2, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2 |
| store i16 %cttz3, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2 |
| store i16 %cttz4, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2 |
| store i16 %cttz5, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2 |
| store i16 %cttz6, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2 |
| store i16 %cttz7, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2 |
| ret void |
| } |
| |
| define void @cttz_undef_16i16() #0 { |
| ; SSE-LABEL: @cttz_undef_16i16( |
| ; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, <8 x i16>* bitcast ([16 x i16]* @src16 to <8 x i16>*), align 2 |
| ; SSE-NEXT: [[TMP2:%.*]] = load <8 x i16>, <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8) to <8 x i16>*), align 2 |
| ; SSE-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[TMP1]], i1 true) |
| ; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.cttz.v8i16(<8 x i16> [[TMP2]], i1 true) |
| ; SSE-NEXT: store <8 x i16> [[TMP3]], <8 x i16>* bitcast ([16 x i16]* @dst16 to <8 x i16>*), align 2 |
| ; SSE-NEXT: store <8 x i16> [[TMP4]], <8 x i16>* bitcast (i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8) to <8 x i16>*), align 2 |
| ; SSE-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_undef_16i16( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, <16 x i16>* bitcast ([16 x i16]* @src16 to <16 x i16>*), align 2 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.cttz.v16i16(<16 x i16> [[TMP1]], i1 true) |
| ; AVX-NEXT: store <16 x i16> [[TMP2]], <16 x i16>* bitcast ([16 x i16]* @dst16 to <16 x i16>*), align 2 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 0), align 2 |
| %ld1 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 1), align 2 |
| %ld2 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 2), align 2 |
| %ld3 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 3), align 2 |
| %ld4 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 4), align 2 |
| %ld5 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 5), align 2 |
| %ld6 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 6), align 2 |
| %ld7 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 7), align 2 |
| %ld8 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 8), align 2 |
| %ld9 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 9), align 2 |
| %ld10 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 10), align 2 |
| %ld11 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 11), align 2 |
| %ld12 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 12), align 2 |
| %ld13 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 13), align 2 |
| %ld14 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 14), align 2 |
| %ld15 = load i16, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @src16, i16 0, i64 15), align 2 |
| %cttz0 = call i16 @llvm.cttz.i16(i16 %ld0, i1 -1) |
| %cttz1 = call i16 @llvm.cttz.i16(i16 %ld1, i1 -1) |
| %cttz2 = call i16 @llvm.cttz.i16(i16 %ld2, i1 -1) |
| %cttz3 = call i16 @llvm.cttz.i16(i16 %ld3, i1 -1) |
| %cttz4 = call i16 @llvm.cttz.i16(i16 %ld4, i1 -1) |
| %cttz5 = call i16 @llvm.cttz.i16(i16 %ld5, i1 -1) |
| %cttz6 = call i16 @llvm.cttz.i16(i16 %ld6, i1 -1) |
| %cttz7 = call i16 @llvm.cttz.i16(i16 %ld7, i1 -1) |
| %cttz8 = call i16 @llvm.cttz.i16(i16 %ld8, i1 -1) |
| %cttz9 = call i16 @llvm.cttz.i16(i16 %ld9, i1 -1) |
| %cttz10 = call i16 @llvm.cttz.i16(i16 %ld10, i1 -1) |
| %cttz11 = call i16 @llvm.cttz.i16(i16 %ld11, i1 -1) |
| %cttz12 = call i16 @llvm.cttz.i16(i16 %ld12, i1 -1) |
| %cttz13 = call i16 @llvm.cttz.i16(i16 %ld13, i1 -1) |
| %cttz14 = call i16 @llvm.cttz.i16(i16 %ld14, i1 -1) |
| %cttz15 = call i16 @llvm.cttz.i16(i16 %ld15, i1 -1) |
| store i16 %cttz0 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 0), align 2 |
| store i16 %cttz1 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 1), align 2 |
| store i16 %cttz2 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 2), align 2 |
| store i16 %cttz3 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 3), align 2 |
| store i16 %cttz4 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 4), align 2 |
| store i16 %cttz5 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 5), align 2 |
| store i16 %cttz6 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 6), align 2 |
| store i16 %cttz7 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 7), align 2 |
| store i16 %cttz8 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 8), align 2 |
| store i16 %cttz9 , i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 9), align 2 |
| store i16 %cttz10, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 10), align 2 |
| store i16 %cttz11, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 11), align 2 |
| store i16 %cttz12, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 12), align 2 |
| store i16 %cttz13, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 13), align 2 |
| store i16 %cttz14, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 14), align 2 |
| store i16 %cttz15, i16* getelementptr inbounds ([16 x i16], [16 x i16]* @dst16, i16 0, i64 15), align 2 |
| ret void |
| } |
| |
| define void @cttz_undef_16i8() #0 { |
| ; CHECK-LABEL: @cttz_undef_16i8( |
| ; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, <16 x i8>* bitcast ([32 x i8]* @src8 to <16 x i8>*), align 1 |
| ; CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.cttz.v16i8(<16 x i8> [[TMP1]], i1 true) |
| ; CHECK-NEXT: store <16 x i8> [[TMP2]], <16 x i8>* bitcast ([32 x i8]* @dst8 to <16 x i8>*), align 1 |
| ; CHECK-NEXT: ret void |
| ; |
| %ld0 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 0), align 1 |
| %ld1 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 1), align 1 |
| %ld2 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 2), align 1 |
| %ld3 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 3), align 1 |
| %ld4 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 4), align 1 |
| %ld5 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 5), align 1 |
| %ld6 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 6), align 1 |
| %ld7 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 7), align 1 |
| %ld8 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 8), align 1 |
| %ld9 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 9), align 1 |
| %ld10 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 10), align 1 |
| %ld11 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 11), align 1 |
| %ld12 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 12), align 1 |
| %ld13 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 13), align 1 |
| %ld14 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 14), align 1 |
| %ld15 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 15), align 1 |
| %cttz0 = call i8 @llvm.cttz.i8(i8 %ld0, i1 -1) |
| %cttz1 = call i8 @llvm.cttz.i8(i8 %ld1, i1 -1) |
| %cttz2 = call i8 @llvm.cttz.i8(i8 %ld2, i1 -1) |
| %cttz3 = call i8 @llvm.cttz.i8(i8 %ld3, i1 -1) |
| %cttz4 = call i8 @llvm.cttz.i8(i8 %ld4, i1 -1) |
| %cttz5 = call i8 @llvm.cttz.i8(i8 %ld5, i1 -1) |
| %cttz6 = call i8 @llvm.cttz.i8(i8 %ld6, i1 -1) |
| %cttz7 = call i8 @llvm.cttz.i8(i8 %ld7, i1 -1) |
| %cttz8 = call i8 @llvm.cttz.i8(i8 %ld8, i1 -1) |
| %cttz9 = call i8 @llvm.cttz.i8(i8 %ld9, i1 -1) |
| %cttz10 = call i8 @llvm.cttz.i8(i8 %ld10, i1 -1) |
| %cttz11 = call i8 @llvm.cttz.i8(i8 %ld11, i1 -1) |
| %cttz12 = call i8 @llvm.cttz.i8(i8 %ld12, i1 -1) |
| %cttz13 = call i8 @llvm.cttz.i8(i8 %ld13, i1 -1) |
| %cttz14 = call i8 @llvm.cttz.i8(i8 %ld14, i1 -1) |
| %cttz15 = call i8 @llvm.cttz.i8(i8 %ld15, i1 -1) |
| store i8 %cttz0 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 0), align 1 |
| store i8 %cttz1 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 1), align 1 |
| store i8 %cttz2 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 2), align 1 |
| store i8 %cttz3 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 3), align 1 |
| store i8 %cttz4 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 4), align 1 |
| store i8 %cttz5 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 5), align 1 |
| store i8 %cttz6 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 6), align 1 |
| store i8 %cttz7 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 7), align 1 |
| store i8 %cttz8 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 8), align 1 |
| store i8 %cttz9 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 9), align 1 |
| store i8 %cttz10, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 10), align 1 |
| store i8 %cttz11, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 11), align 1 |
| store i8 %cttz12, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 12), align 1 |
| store i8 %cttz13, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 13), align 1 |
| store i8 %cttz14, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 14), align 1 |
| store i8 %cttz15, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 15), align 1 |
| ret void |
| } |
| |
| define void @cttz_undef_32i8() #0 { |
| ; SSE-LABEL: @cttz_undef_32i8( |
| ; SSE-NEXT: [[TMP1:%.*]] = load <16 x i8>, <16 x i8>* bitcast ([32 x i8]* @src8 to <16 x i8>*), align 1 |
| ; SSE-NEXT: [[TMP2:%.*]] = load <16 x i8>, <16 x i8>* bitcast (i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 16) to <16 x i8>*), align 1 |
| ; SSE-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.cttz.v16i8(<16 x i8> [[TMP1]], i1 true) |
| ; SSE-NEXT: [[TMP4:%.*]] = call <16 x i8> @llvm.cttz.v16i8(<16 x i8> [[TMP2]], i1 true) |
| ; SSE-NEXT: store <16 x i8> [[TMP3]], <16 x i8>* bitcast ([32 x i8]* @dst8 to <16 x i8>*), align 1 |
| ; SSE-NEXT: store <16 x i8> [[TMP4]], <16 x i8>* bitcast (i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 16) to <16 x i8>*), align 1 |
| ; SSE-NEXT: ret void |
| ; |
| ; AVX-LABEL: @cttz_undef_32i8( |
| ; AVX-NEXT: [[TMP1:%.*]] = load <32 x i8>, <32 x i8>* bitcast ([32 x i8]* @src8 to <32 x i8>*), align 1 |
| ; AVX-NEXT: [[TMP2:%.*]] = call <32 x i8> @llvm.cttz.v32i8(<32 x i8> [[TMP1]], i1 true) |
| ; AVX-NEXT: store <32 x i8> [[TMP2]], <32 x i8>* bitcast ([32 x i8]* @dst8 to <32 x i8>*), align 1 |
| ; AVX-NEXT: ret void |
| ; |
| %ld0 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 0), align 1 |
| %ld1 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 1), align 1 |
| %ld2 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 2), align 1 |
| %ld3 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 3), align 1 |
| %ld4 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 4), align 1 |
| %ld5 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 5), align 1 |
| %ld6 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 6), align 1 |
| %ld7 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 7), align 1 |
| %ld8 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 8), align 1 |
| %ld9 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 9), align 1 |
| %ld10 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 10), align 1 |
| %ld11 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 11), align 1 |
| %ld12 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 12), align 1 |
| %ld13 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 13), align 1 |
| %ld14 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 14), align 1 |
| %ld15 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 15), align 1 |
| %ld16 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 16), align 1 |
| %ld17 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 17), align 1 |
| %ld18 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 18), align 1 |
| %ld19 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 19), align 1 |
| %ld20 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 20), align 1 |
| %ld21 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 21), align 1 |
| %ld22 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 22), align 1 |
| %ld23 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 23), align 1 |
| %ld24 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 24), align 1 |
| %ld25 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 25), align 1 |
| %ld26 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 26), align 1 |
| %ld27 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 27), align 1 |
| %ld28 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 28), align 1 |
| %ld29 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 29), align 1 |
| %ld30 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 30), align 1 |
| %ld31 = load i8, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @src8, i8 0, i64 31), align 1 |
| %cttz0 = call i8 @llvm.cttz.i8(i8 %ld0, i1 -1) |
| %cttz1 = call i8 @llvm.cttz.i8(i8 %ld1, i1 -1) |
| %cttz2 = call i8 @llvm.cttz.i8(i8 %ld2, i1 -1) |
| %cttz3 = call i8 @llvm.cttz.i8(i8 %ld3, i1 -1) |
| %cttz4 = call i8 @llvm.cttz.i8(i8 %ld4, i1 -1) |
| %cttz5 = call i8 @llvm.cttz.i8(i8 %ld5, i1 -1) |
| %cttz6 = call i8 @llvm.cttz.i8(i8 %ld6, i1 -1) |
| %cttz7 = call i8 @llvm.cttz.i8(i8 %ld7, i1 -1) |
| %cttz8 = call i8 @llvm.cttz.i8(i8 %ld8, i1 -1) |
| %cttz9 = call i8 @llvm.cttz.i8(i8 %ld9, i1 -1) |
| %cttz10 = call i8 @llvm.cttz.i8(i8 %ld10, i1 -1) |
| %cttz11 = call i8 @llvm.cttz.i8(i8 %ld11, i1 -1) |
| %cttz12 = call i8 @llvm.cttz.i8(i8 %ld12, i1 -1) |
| %cttz13 = call i8 @llvm.cttz.i8(i8 %ld13, i1 -1) |
| %cttz14 = call i8 @llvm.cttz.i8(i8 %ld14, i1 -1) |
| %cttz15 = call i8 @llvm.cttz.i8(i8 %ld15, i1 -1) |
| %cttz16 = call i8 @llvm.cttz.i8(i8 %ld16, i1 -1) |
| %cttz17 = call i8 @llvm.cttz.i8(i8 %ld17, i1 -1) |
| %cttz18 = call i8 @llvm.cttz.i8(i8 %ld18, i1 -1) |
| %cttz19 = call i8 @llvm.cttz.i8(i8 %ld19, i1 -1) |
| %cttz20 = call i8 @llvm.cttz.i8(i8 %ld20, i1 -1) |
| %cttz21 = call i8 @llvm.cttz.i8(i8 %ld21, i1 -1) |
| %cttz22 = call i8 @llvm.cttz.i8(i8 %ld22, i1 -1) |
| %cttz23 = call i8 @llvm.cttz.i8(i8 %ld23, i1 -1) |
| %cttz24 = call i8 @llvm.cttz.i8(i8 %ld24, i1 -1) |
| %cttz25 = call i8 @llvm.cttz.i8(i8 %ld25, i1 -1) |
| %cttz26 = call i8 @llvm.cttz.i8(i8 %ld26, i1 -1) |
| %cttz27 = call i8 @llvm.cttz.i8(i8 %ld27, i1 -1) |
| %cttz28 = call i8 @llvm.cttz.i8(i8 %ld28, i1 -1) |
| %cttz29 = call i8 @llvm.cttz.i8(i8 %ld29, i1 -1) |
| %cttz30 = call i8 @llvm.cttz.i8(i8 %ld30, i1 -1) |
| %cttz31 = call i8 @llvm.cttz.i8(i8 %ld31, i1 -1) |
| store i8 %cttz0 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 0), align 1 |
| store i8 %cttz1 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 1), align 1 |
| store i8 %cttz2 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 2), align 1 |
| store i8 %cttz3 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 3), align 1 |
| store i8 %cttz4 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 4), align 1 |
| store i8 %cttz5 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 5), align 1 |
| store i8 %cttz6 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 6), align 1 |
| store i8 %cttz7 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 7), align 1 |
| store i8 %cttz8 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 8), align 1 |
| store i8 %cttz9 , i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 9), align 1 |
| store i8 %cttz10, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 10), align 1 |
| store i8 %cttz11, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 11), align 1 |
| store i8 %cttz12, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 12), align 1 |
| store i8 %cttz13, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 13), align 1 |
| store i8 %cttz14, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 14), align 1 |
| store i8 %cttz15, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 15), align 1 |
| store i8 %cttz16, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 16), align 1 |
| store i8 %cttz17, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 17), align 1 |
| store i8 %cttz18, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 18), align 1 |
| store i8 %cttz19, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 19), align 1 |
| store i8 %cttz20, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 20), align 1 |
| store i8 %cttz21, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 21), align 1 |
| store i8 %cttz22, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 22), align 1 |
| store i8 %cttz23, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 23), align 1 |
| store i8 %cttz24, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 24), align 1 |
| store i8 %cttz25, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 25), align 1 |
| store i8 %cttz26, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 26), align 1 |
| store i8 %cttz27, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 27), align 1 |
| store i8 %cttz28, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 28), align 1 |
| store i8 %cttz29, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 29), align 1 |
| store i8 %cttz30, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 30), align 1 |
| store i8 %cttz31, i8* getelementptr inbounds ([32 x i8], [32 x i8]* @dst8, i8 0, i64 31), align 1 |
| ret void |
| } |
| |
| attributes #0 = { nounwind } |
