; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -mtriple=x86_64-unknown -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=bdver2 -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=XOP
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=bdver4 -passes=slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=XOP
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.bitreverse.i64(i64)
declare i32 @llvm.bitreverse.i32(i32)
declare i16 @llvm.bitreverse.i16(i16)
declare i8 @llvm.bitreverse.i8(i8)
define void @bitreverse_2i64() #0 {
; CHECK-LABEL: @bitreverse_2i64(
; CHECK-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @src64, align 8
; CHECK-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.bitreverse.v2i64(<2 x i64> [[TMP1]])
; CHECK-NEXT: store <2 x i64> [[TMP2]], ptr @dst64, align 8
; CHECK-NEXT: ret void
;
%ld0 = load i64, ptr @src64, align 8
%ld1 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i32 0, i64 1), align 8
%bitreverse0 = call i64 @llvm.bitreverse.i64(i64 %ld0)
%bitreverse1 = call i64 @llvm.bitreverse.i64(i64 %ld1)
store i64 %bitreverse0, ptr @dst64, align 8
store i64 %bitreverse1, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i32 0, i64 1), align 8
ret void
}
define void @bitreverse_4i64() #0 {
; SSE-LABEL: @bitreverse_4i64(
; SSE-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @src64, align 4
; SSE-NEXT: [[TMP2:%.*]] = call <2 x i64> @llvm.bitreverse.v2i64(<2 x i64> [[TMP1]])
; SSE-NEXT: store <2 x i64> [[TMP2]], ptr @dst64, align 4
; SSE-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 2), align 4
; SSE-NEXT: [[TMP4:%.*]] = call <2 x i64> @llvm.bitreverse.v2i64(<2 x i64> [[TMP3]])
; SSE-NEXT: store <2 x i64> [[TMP4]], ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 2), align 4
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_4i64(
; AVX-NEXT: [[TMP1:%.*]] = load <4 x i64>, ptr @src64, align 4
; AVX-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.bitreverse.v4i64(<4 x i64> [[TMP1]])
; AVX-NEXT: store <4 x i64> [[TMP2]], ptr @dst64, align 4
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_4i64(
; XOP-NEXT: [[TMP1:%.*]] = load <4 x i64>, ptr @src64, align 4
; XOP-NEXT: [[TMP2:%.*]] = call <4 x i64> @llvm.bitreverse.v4i64(<4 x i64> [[TMP1]])
; XOP-NEXT: store <4 x i64> [[TMP2]], ptr @dst64, align 4
; XOP-NEXT: ret void
;
%ld0 = load i64, ptr @src64, align 4
%ld1 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 1), align 4
%ld2 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 2), align 4
%ld3 = load i64, ptr getelementptr inbounds ([4 x i64], ptr @src64, i64 0, i64 3), align 4
%bitreverse0 = call i64 @llvm.bitreverse.i64(i64 %ld0)
%bitreverse1 = call i64 @llvm.bitreverse.i64(i64 %ld1)
%bitreverse2 = call i64 @llvm.bitreverse.i64(i64 %ld2)
%bitreverse3 = call i64 @llvm.bitreverse.i64(i64 %ld3)
store i64 %bitreverse0, ptr @dst64, align 4
store i64 %bitreverse1, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 1), align 4
store i64 %bitreverse2, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 2), align 4
store i64 %bitreverse3, ptr getelementptr inbounds ([4 x i64], ptr @dst64, i64 0, i64 3), align 4
ret void
}
define void @bitreverse_4i32() #0 {
; CHECK-LABEL: @bitreverse_4i32(
; CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @src32, align 4
; CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.bitreverse.v4i32(<4 x i32> [[TMP1]])
; CHECK-NEXT: store <4 x i32> [[TMP2]], ptr @dst32, align 4
; CHECK-NEXT: ret void
;
%ld0 = load i32, ptr @src32, align 4
%ld1 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 1), align 4
%ld2 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 2), align 4
%ld3 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 3), align 4
%bitreverse0 = call i32 @llvm.bitreverse.i32(i32 %ld0)
%bitreverse1 = call i32 @llvm.bitreverse.i32(i32 %ld1)
%bitreverse2 = call i32 @llvm.bitreverse.i32(i32 %ld2)
%bitreverse3 = call i32 @llvm.bitreverse.i32(i32 %ld3)
store i32 %bitreverse0, ptr @dst32, align 4
store i32 %bitreverse1, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 1), align 4
store i32 %bitreverse2, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 2), align 4
store i32 %bitreverse3, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 3), align 4
ret void
}
define void @bitreverse_8i32() #0 {
; SSE-LABEL: @bitreverse_8i32(
; SSE-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @src32, align 2
; SSE-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.bitreverse.v4i32(<4 x i32> [[TMP1]])
; SSE-NEXT: store <4 x i32> [[TMP2]], ptr @dst32, align 2
; SSE-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 4), align 2
; SSE-NEXT: [[TMP4:%.*]] = call <4 x i32> @llvm.bitreverse.v4i32(<4 x i32> [[TMP3]])
; SSE-NEXT: store <4 x i32> [[TMP4]], ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 4), align 2
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_8i32(
; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, ptr @src32, align 2
; AVX-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.bitreverse.v8i32(<8 x i32> [[TMP1]])
; AVX-NEXT: store <8 x i32> [[TMP2]], ptr @dst32, align 2
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_8i32(
; XOP-NEXT: [[TMP1:%.*]] = load <8 x i32>, ptr @src32, align 2
; XOP-NEXT: [[TMP2:%.*]] = call <8 x i32> @llvm.bitreverse.v8i32(<8 x i32> [[TMP1]])
; XOP-NEXT: store <8 x i32> [[TMP2]], ptr @dst32, align 2
; XOP-NEXT: ret void
;
%ld0 = load i32, ptr @src32, align 2
%ld1 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 1), align 2
%ld2 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 2), align 2
%ld3 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 3), align 2
%ld4 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 4), align 2
%ld5 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 5), align 2
%ld6 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 6), align 2
%ld7 = load i32, ptr getelementptr inbounds ([8 x i32], ptr @src32, i32 0, i64 7), align 2
%bitreverse0 = call i32 @llvm.bitreverse.i32(i32 %ld0)
%bitreverse1 = call i32 @llvm.bitreverse.i32(i32 %ld1)
%bitreverse2 = call i32 @llvm.bitreverse.i32(i32 %ld2)
%bitreverse3 = call i32 @llvm.bitreverse.i32(i32 %ld3)
%bitreverse4 = call i32 @llvm.bitreverse.i32(i32 %ld4)
%bitreverse5 = call i32 @llvm.bitreverse.i32(i32 %ld5)
%bitreverse6 = call i32 @llvm.bitreverse.i32(i32 %ld6)
%bitreverse7 = call i32 @llvm.bitreverse.i32(i32 %ld7)
store i32 %bitreverse0, ptr @dst32, align 2
store i32 %bitreverse1, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 1), align 2
store i32 %bitreverse2, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 2), align 2
store i32 %bitreverse3, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 3), align 2
store i32 %bitreverse4, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 4), align 2
store i32 %bitreverse5, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 5), align 2
store i32 %bitreverse6, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 6), align 2
store i32 %bitreverse7, ptr getelementptr inbounds ([8 x i32], ptr @dst32, i32 0, i64 7), align 2
ret void
}
define void @bitreverse_8i16() #0 {
; CHECK-LABEL: @bitreverse_8i16(
; CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @src16, align 2
; CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.bitreverse.v8i16(<8 x i16> [[TMP1]])
; CHECK-NEXT: store <8 x i16> [[TMP2]], ptr @dst16, align 2
; CHECK-NEXT: ret void
;
%ld0 = load i16, ptr @src16, align 2
%ld1 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 1), align 2
%ld2 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 2), align 2
%ld3 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 3), align 2
%ld4 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 4), align 2
%ld5 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 5), align 2
%ld6 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 6), align 2
%ld7 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 7), align 2
%bitreverse0 = call i16 @llvm.bitreverse.i16(i16 %ld0)
%bitreverse1 = call i16 @llvm.bitreverse.i16(i16 %ld1)
%bitreverse2 = call i16 @llvm.bitreverse.i16(i16 %ld2)
%bitreverse3 = call i16 @llvm.bitreverse.i16(i16 %ld3)
%bitreverse4 = call i16 @llvm.bitreverse.i16(i16 %ld4)
%bitreverse5 = call i16 @llvm.bitreverse.i16(i16 %ld5)
%bitreverse6 = call i16 @llvm.bitreverse.i16(i16 %ld6)
%bitreverse7 = call i16 @llvm.bitreverse.i16(i16 %ld7)
store i16 %bitreverse0, ptr @dst16, align 2
store i16 %bitreverse1, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 1), align 2
store i16 %bitreverse2, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 2), align 2
store i16 %bitreverse3, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 3), align 2
store i16 %bitreverse4, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 4), align 2
store i16 %bitreverse5, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 5), align 2
store i16 %bitreverse6, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 6), align 2
store i16 %bitreverse7, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 7), align 2
ret void
}
define void @bitreverse_16i16() #0 {
; SSE-LABEL: @bitreverse_16i16(
; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @src16, align 2
; SSE-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.bitreverse.v8i16(<8 x i16> [[TMP1]])
; SSE-NEXT: store <8 x i16> [[TMP2]], ptr @dst16, align 2
; SSE-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 8), align 2
; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.bitreverse.v8i16(<8 x i16> [[TMP3]])
; SSE-NEXT: store <8 x i16> [[TMP4]], ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 8), align 2
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_16i16(
; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, ptr @src16, align 2
; AVX-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.bitreverse.v16i16(<16 x i16> [[TMP1]])
; AVX-NEXT: store <16 x i16> [[TMP2]], ptr @dst16, align 2
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_16i16(
; XOP-NEXT: [[TMP1:%.*]] = load <16 x i16>, ptr @src16, align 2
; XOP-NEXT: [[TMP2:%.*]] = call <16 x i16> @llvm.bitreverse.v16i16(<16 x i16> [[TMP1]])
; XOP-NEXT: store <16 x i16> [[TMP2]], ptr @dst16, align 2
; XOP-NEXT: ret void
;
%ld0 = load i16, ptr @src16, align 2
%ld1 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 1), align 2
%ld2 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 2), align 2
%ld3 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 3), align 2
%ld4 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 4), align 2
%ld5 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 5), align 2
%ld6 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 6), align 2
%ld7 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 7), align 2
%ld8 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 8), align 2
%ld9 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 9), align 2
%ld10 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 10), align 2
%ld11 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 11), align 2
%ld12 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 12), align 2
%ld13 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 13), align 2
%ld14 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 14), align 2
%ld15 = load i16, ptr getelementptr inbounds ([16 x i16], ptr @src16, i16 0, i64 15), align 2
%bitreverse0 = call i16 @llvm.bitreverse.i16(i16 %ld0)
%bitreverse1 = call i16 @llvm.bitreverse.i16(i16 %ld1)
%bitreverse2 = call i16 @llvm.bitreverse.i16(i16 %ld2)
%bitreverse3 = call i16 @llvm.bitreverse.i16(i16 %ld3)
%bitreverse4 = call i16 @llvm.bitreverse.i16(i16 %ld4)
%bitreverse5 = call i16 @llvm.bitreverse.i16(i16 %ld5)
%bitreverse6 = call i16 @llvm.bitreverse.i16(i16 %ld6)
%bitreverse7 = call i16 @llvm.bitreverse.i16(i16 %ld7)
%bitreverse8 = call i16 @llvm.bitreverse.i16(i16 %ld8)
%bitreverse9 = call i16 @llvm.bitreverse.i16(i16 %ld9)
%bitreverse10 = call i16 @llvm.bitreverse.i16(i16 %ld10)
%bitreverse11 = call i16 @llvm.bitreverse.i16(i16 %ld11)
%bitreverse12 = call i16 @llvm.bitreverse.i16(i16 %ld12)
%bitreverse13 = call i16 @llvm.bitreverse.i16(i16 %ld13)
%bitreverse14 = call i16 @llvm.bitreverse.i16(i16 %ld14)
%bitreverse15 = call i16 @llvm.bitreverse.i16(i16 %ld15)
store i16 %bitreverse0 , ptr @dst16, align 2
store i16 %bitreverse1 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 1), align 2
store i16 %bitreverse2 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 2), align 2
store i16 %bitreverse3 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 3), align 2
store i16 %bitreverse4 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 4), align 2
store i16 %bitreverse5 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 5), align 2
store i16 %bitreverse6 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 6), align 2
store i16 %bitreverse7 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 7), align 2
store i16 %bitreverse8 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 8), align 2
store i16 %bitreverse9 , ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 9), align 2
store i16 %bitreverse10, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 10), align 2
store i16 %bitreverse11, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 11), align 2
store i16 %bitreverse12, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 12), align 2
store i16 %bitreverse13, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 13), align 2
store i16 %bitreverse14, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 14), align 2
store i16 %bitreverse15, ptr getelementptr inbounds ([16 x i16], ptr @dst16, i16 0, i64 15), align 2
ret void
}
define void @bitreverse_16i8() #0 {
; CHECK-LABEL: @bitreverse_16i8(
; CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @src8, align 1
; CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.bitreverse.v16i8(<16 x i8> [[TMP1]])
; CHECK-NEXT: store <16 x i8> [[TMP2]], ptr @dst8, align 1
; CHECK-NEXT: ret void
;
%ld0 = load i8, ptr @src8, align 1
%ld1 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 1), align 1
%ld2 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 2), align 1
%ld3 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 3), align 1
%ld4 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 4), align 1
%ld5 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 5), align 1
%ld6 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 6), align 1
%ld7 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 7), align 1
%ld8 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 8), align 1
%ld9 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 9), align 1
%ld10 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 10), align 1
%ld11 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 11), align 1
%ld12 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 12), align 1
%ld13 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 13), align 1
%ld14 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 14), align 1
%ld15 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 15), align 1
%bitreverse0 = call i8 @llvm.bitreverse.i8(i8 %ld0)
%bitreverse1 = call i8 @llvm.bitreverse.i8(i8 %ld1)
%bitreverse2 = call i8 @llvm.bitreverse.i8(i8 %ld2)
%bitreverse3 = call i8 @llvm.bitreverse.i8(i8 %ld3)
%bitreverse4 = call i8 @llvm.bitreverse.i8(i8 %ld4)
%bitreverse5 = call i8 @llvm.bitreverse.i8(i8 %ld5)
%bitreverse6 = call i8 @llvm.bitreverse.i8(i8 %ld6)
%bitreverse7 = call i8 @llvm.bitreverse.i8(i8 %ld7)
%bitreverse8 = call i8 @llvm.bitreverse.i8(i8 %ld8)
%bitreverse9 = call i8 @llvm.bitreverse.i8(i8 %ld9)
%bitreverse10 = call i8 @llvm.bitreverse.i8(i8 %ld10)
%bitreverse11 = call i8 @llvm.bitreverse.i8(i8 %ld11)
%bitreverse12 = call i8 @llvm.bitreverse.i8(i8 %ld12)
%bitreverse13 = call i8 @llvm.bitreverse.i8(i8 %ld13)
%bitreverse14 = call i8 @llvm.bitreverse.i8(i8 %ld14)
%bitreverse15 = call i8 @llvm.bitreverse.i8(i8 %ld15)
store i8 %bitreverse0 , ptr @dst8, align 1
store i8 %bitreverse1 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 1), align 1
store i8 %bitreverse2 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 2), align 1
store i8 %bitreverse3 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 3), align 1
store i8 %bitreverse4 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 4), align 1
store i8 %bitreverse5 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 5), align 1
store i8 %bitreverse6 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 6), align 1
store i8 %bitreverse7 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 7), align 1
store i8 %bitreverse8 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 8), align 1
store i8 %bitreverse9 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 9), align 1
store i8 %bitreverse10, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 10), align 1
store i8 %bitreverse11, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 11), align 1
store i8 %bitreverse12, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 12), align 1
store i8 %bitreverse13, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 13), align 1
store i8 %bitreverse14, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 14), align 1
store i8 %bitreverse15, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 15), align 1
ret void
}
define void @bitreverse_32i8() #0 {
; SSE-LABEL: @bitreverse_32i8(
; SSE-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @src8, align 1
; SSE-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.bitreverse.v16i8(<16 x i8> [[TMP1]])
; SSE-NEXT: store <16 x i8> [[TMP2]], ptr @dst8, align 1
; SSE-NEXT: [[TMP3:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 16), align 1
; SSE-NEXT: [[TMP4:%.*]] = call <16 x i8> @llvm.bitreverse.v16i8(<16 x i8> [[TMP3]])
; SSE-NEXT: store <16 x i8> [[TMP4]], ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 16), align 1
; SSE-NEXT: ret void
;
; AVX-LABEL: @bitreverse_32i8(
; AVX-NEXT: [[TMP1:%.*]] = load <32 x i8>, ptr @src8, align 1
; AVX-NEXT: [[TMP2:%.*]] = call <32 x i8> @llvm.bitreverse.v32i8(<32 x i8> [[TMP1]])
; AVX-NEXT: store <32 x i8> [[TMP2]], ptr @dst8, align 1
; AVX-NEXT: ret void
;
; XOP-LABEL: @bitreverse_32i8(
; XOP-NEXT: [[TMP1:%.*]] = load <32 x i8>, ptr @src8, align 1
; XOP-NEXT: [[TMP2:%.*]] = call <32 x i8> @llvm.bitreverse.v32i8(<32 x i8> [[TMP1]])
; XOP-NEXT: store <32 x i8> [[TMP2]], ptr @dst8, align 1
; XOP-NEXT: ret void
;
%ld0 = load i8, ptr @src8, align 1
%ld1 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 1), align 1
%ld2 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 2), align 1
%ld3 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 3), align 1
%ld4 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 4), align 1
%ld5 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 5), align 1
%ld6 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 6), align 1
%ld7 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 7), align 1
%ld8 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 8), align 1
%ld9 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 9), align 1
%ld10 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 10), align 1
%ld11 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 11), align 1
%ld12 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 12), align 1
%ld13 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 13), align 1
%ld14 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 14), align 1
%ld15 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 15), align 1
%ld16 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 16), align 1
%ld17 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 17), align 1
%ld18 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 18), align 1
%ld19 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 19), align 1
%ld20 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 20), align 1
%ld21 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 21), align 1
%ld22 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 22), align 1
%ld23 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 23), align 1
%ld24 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 24), align 1
%ld25 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 25), align 1
%ld26 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 26), align 1
%ld27 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 27), align 1
%ld28 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 28), align 1
%ld29 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 29), align 1
%ld30 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 30), align 1
%ld31 = load i8, ptr getelementptr inbounds ([32 x i8], ptr @src8, i8 0, i64 31), align 1
%bitreverse0 = call i8 @llvm.bitreverse.i8(i8 %ld0)
%bitreverse1 = call i8 @llvm.bitreverse.i8(i8 %ld1)
%bitreverse2 = call i8 @llvm.bitreverse.i8(i8 %ld2)
%bitreverse3 = call i8 @llvm.bitreverse.i8(i8 %ld3)
%bitreverse4 = call i8 @llvm.bitreverse.i8(i8 %ld4)
%bitreverse5 = call i8 @llvm.bitreverse.i8(i8 %ld5)
%bitreverse6 = call i8 @llvm.bitreverse.i8(i8 %ld6)
%bitreverse7 = call i8 @llvm.bitreverse.i8(i8 %ld7)
%bitreverse8 = call i8 @llvm.bitreverse.i8(i8 %ld8)
%bitreverse9 = call i8 @llvm.bitreverse.i8(i8 %ld9)
%bitreverse10 = call i8 @llvm.bitreverse.i8(i8 %ld10)
%bitreverse11 = call i8 @llvm.bitreverse.i8(i8 %ld11)
%bitreverse12 = call i8 @llvm.bitreverse.i8(i8 %ld12)
%bitreverse13 = call i8 @llvm.bitreverse.i8(i8 %ld13)
%bitreverse14 = call i8 @llvm.bitreverse.i8(i8 %ld14)
%bitreverse15 = call i8 @llvm.bitreverse.i8(i8 %ld15)
%bitreverse16 = call i8 @llvm.bitreverse.i8(i8 %ld16)
%bitreverse17 = call i8 @llvm.bitreverse.i8(i8 %ld17)
%bitreverse18 = call i8 @llvm.bitreverse.i8(i8 %ld18)
%bitreverse19 = call i8 @llvm.bitreverse.i8(i8 %ld19)
%bitreverse20 = call i8 @llvm.bitreverse.i8(i8 %ld20)
%bitreverse21 = call i8 @llvm.bitreverse.i8(i8 %ld21)
%bitreverse22 = call i8 @llvm.bitreverse.i8(i8 %ld22)
%bitreverse23 = call i8 @llvm.bitreverse.i8(i8 %ld23)
%bitreverse24 = call i8 @llvm.bitreverse.i8(i8 %ld24)
%bitreverse25 = call i8 @llvm.bitreverse.i8(i8 %ld25)
%bitreverse26 = call i8 @llvm.bitreverse.i8(i8 %ld26)
%bitreverse27 = call i8 @llvm.bitreverse.i8(i8 %ld27)
%bitreverse28 = call i8 @llvm.bitreverse.i8(i8 %ld28)
%bitreverse29 = call i8 @llvm.bitreverse.i8(i8 %ld29)
%bitreverse30 = call i8 @llvm.bitreverse.i8(i8 %ld30)
%bitreverse31 = call i8 @llvm.bitreverse.i8(i8 %ld31)
store i8 %bitreverse0 , ptr @dst8, align 1
store i8 %bitreverse1 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 1), align 1
store i8 %bitreverse2 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 2), align 1
store i8 %bitreverse3 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 3), align 1
store i8 %bitreverse4 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 4), align 1
store i8 %bitreverse5 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 5), align 1
store i8 %bitreverse6 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 6), align 1
store i8 %bitreverse7 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 7), align 1
store i8 %bitreverse8 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 8), align 1
store i8 %bitreverse9 , ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 9), align 1
store i8 %bitreverse10, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 10), align 1
store i8 %bitreverse11, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 11), align 1
store i8 %bitreverse12, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 12), align 1
store i8 %bitreverse13, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 13), align 1
store i8 %bitreverse14, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 14), align 1
store i8 %bitreverse15, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 15), align 1
store i8 %bitreverse16, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 16), align 1
store i8 %bitreverse17, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 17), align 1
store i8 %bitreverse18, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 18), align 1
store i8 %bitreverse19, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 19), align 1
store i8 %bitreverse20, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 20), align 1
store i8 %bitreverse21, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 21), align 1
store i8 %bitreverse22, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 22), align 1
store i8 %bitreverse23, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 23), align 1
store i8 %bitreverse24, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 24), align 1
store i8 %bitreverse25, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 25), align 1
store i8 %bitreverse26, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 26), align 1
store i8 %bitreverse27, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 27), align 1
store i8 %bitreverse28, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 28), align 1
store i8 %bitreverse29, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 29), align 1
store i8 %bitreverse30, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 30), align 1
store i8 %bitreverse31, ptr getelementptr inbounds ([32 x i8], ptr @dst8, i8 0, i64 31), align 1
ret void
}
attributes #0 = { nounwind }
Codebase Browser
llvm