[llvm] r282081 - [AVX512] Fix return types on int_x86_avx512_gatherXXX_di intrinsics
Cameron McInally via llvm-commits
llvm-commits at lists.llvm.org
Wed Sep 21 09:06:10 PDT 2016
Author: mcinally
Date: Wed Sep 21 11:06:10 2016
New Revision: 282081
URL: http://llvm.org/viewvc/llvm-project?rev=282081&view=rev
Log:
[AVX512] Fix return types on int_x86_avx512_gatherXXX_di intrinsics
The return type should match the pass through vector type.
Differential Revision: https://reviews.llvm.org/D24744
Modified:
llvm/trunk/include/llvm/IR/IntrinsicsX86.td
llvm/trunk/test/CodeGen/X86/avx512-gather-scatter-intrin.ll
Modified: llvm/trunk/include/llvm/IR/IntrinsicsX86.td
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/IR/IntrinsicsX86.td?rev=282081&r1=282080&r2=282081&view=diff
==============================================================================
--- llvm/trunk/include/llvm/IR/IntrinsicsX86.td (original)
+++ llvm/trunk/include/llvm/IR/IntrinsicsX86.td Wed Sep 21 11:06:10 2016
@@ -5819,7 +5819,7 @@ let TargetPrefix = "x86" in {
def int_x86_avx512_gather3div2_di :
GCCBuiltin<"__builtin_ia32_gather3div2di">,
- Intrinsic<[llvm_v4i32_ty],
+ Intrinsic<[llvm_v2i64_ty],
[llvm_v2i64_ty, llvm_ptr_ty, llvm_v2i64_ty, llvm_i8_ty, llvm_i32_ty],
[IntrReadMem, IntrArgMemOnly]>;
@@ -5831,7 +5831,7 @@ let TargetPrefix = "x86" in {
def int_x86_avx512_gather3div4_di :
GCCBuiltin<"__builtin_ia32_gather3div4di">,
- Intrinsic<[llvm_v8i32_ty],
+ Intrinsic<[llvm_v4i64_ty],
[llvm_v4i64_ty, llvm_ptr_ty, llvm_v4i64_ty, llvm_i8_ty, llvm_i32_ty],
[IntrReadMem, IntrArgMemOnly]>;
@@ -5867,7 +5867,7 @@ let TargetPrefix = "x86" in {
def int_x86_avx512_gather3siv2_di :
GCCBuiltin<"__builtin_ia32_gather3siv2di">,
- Intrinsic<[llvm_v4i32_ty],
+ Intrinsic<[llvm_v2i64_ty],
[llvm_v2i64_ty, llvm_ptr_ty, llvm_v4i32_ty, llvm_i8_ty, llvm_i32_ty],
[IntrReadMem, IntrArgMemOnly]>;
@@ -5879,7 +5879,7 @@ let TargetPrefix = "x86" in {
def int_x86_avx512_gather3siv4_di :
GCCBuiltin<"__builtin_ia32_gather3siv4di">,
- Intrinsic<[llvm_v8i32_ty],
+ Intrinsic<[llvm_v4i64_ty],
[llvm_v4i64_ty, llvm_ptr_ty, llvm_v4i32_ty, llvm_i8_ty, llvm_i32_ty],
[IntrReadMem, IntrArgMemOnly]>;
Modified: llvm/trunk/test/CodeGen/X86/avx512-gather-scatter-intrin.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-gather-scatter-intrin.ll?rev=282081&r1=282080&r2=282081&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-gather-scatter-intrin.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-gather-scatter-intrin.ll Wed Sep 21 11:06:10 2016
@@ -293,19 +293,19 @@ define <2 x double>@test_int_x86_avx512_
ret <2 x double> %res2
}
-declare <4 x i32> @llvm.x86.avx512.gather3div2.di(<2 x i64>, i8*, <2 x i64>, i8, i32)
+declare <2 x i64> @llvm.x86.avx512.gather3div2.di(<2 x i64>, i8*, <2 x i64>, i8, i32)
-define <4 x i32>@test_int_x86_avx512_gather3div2_di(<2 x i64> %x0, i8* %x1, <2 x i64> %x2, i8 %x3) {
+define <2 x i64>@test_int_x86_avx512_gather3div2_di(<2 x i64> %x0, i8* %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_gather3div2_di:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovb %esi, %k1
; CHECK-NEXT: vpgatherqq (%rdi,%xmm1,8), %xmm0 {%k1}
-; CHECK-NEXT: vpaddd %xmm0, %xmm0, %xmm0
+; CHECK-NEXT: vpaddq %xmm0, %xmm0, %xmm0
; CHECK-NEXT: retq
- %res = call <4 x i32> @llvm.x86.avx512.gather3div2.di(<2 x i64> %x0, i8* %x1, <2 x i64> %x2, i8 %x3, i32 8)
- %res1 = call <4 x i32> @llvm.x86.avx512.gather3div2.di(<2 x i64> %x0, i8* %x1, <2 x i64> %x2, i8 %x3, i32 8)
- %res2 = add <4 x i32> %res, %res1
- ret <4 x i32> %res2
+ %res = call <2 x i64> @llvm.x86.avx512.gather3div2.di(<2 x i64> %x0, i8* %x1, <2 x i64> %x2, i8 %x3, i32 8)
+ %res1 = call <2 x i64> @llvm.x86.avx512.gather3div2.di(<2 x i64> %x0, i8* %x1, <2 x i64> %x2, i8 %x3, i32 8)
+ %res2 = add <2 x i64> %res, %res1
+ ret <2 x i64> %res2
}
declare <4 x double> @llvm.x86.avx512.gather3div4.df(<4 x double>, i8*, <4 x i64>, i8, i32)
@@ -326,9 +326,9 @@ define <4 x double>@test_int_x86_avx512_
ret <4 x double> %res2
}
-declare <8 x i32> @llvm.x86.avx512.gather3div4.di(<4 x i64>, i8*, <4 x i64>, i8, i32)
+declare <4 x i64> @llvm.x86.avx512.gather3div4.di(<4 x i64>, i8*, <4 x i64>, i8, i32)
-define <8 x i32>@test_int_x86_avx512_gather3div4_di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 %x3) {
+define <4 x i64>@test_int_x86_avx512_gather3div4_di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_gather3div4_di:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovb %esi, %k1
@@ -336,12 +336,12 @@ define <8 x i32>@test_int_x86_avx512_gat
; CHECK-NEXT: vpgatherqq (%rdi,%ymm1,8), %ymm2 {%k1}
; CHECK-NEXT: kxnorw %k0, %k0, %k1
; CHECK-NEXT: vpgatherqq (%rdi,%ymm1,8), %ymm0 {%k1}
-; CHECK-NEXT: vpaddd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT: vpaddq %ymm0, %ymm2, %ymm0
; CHECK-NEXT: retq
- %res = call <8 x i32> @llvm.x86.avx512.gather3div4.di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 %x3, i32 8)
- %res1 = call <8 x i32> @llvm.x86.avx512.gather3div4.di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 -1, i32 8)
- %res2 = add <8 x i32> %res, %res1
- ret <8 x i32> %res2
+ %res = call <4 x i64> @llvm.x86.avx512.gather3div4.di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 %x3, i32 8)
+ %res1 = call <4 x i64> @llvm.x86.avx512.gather3div4.di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 -1, i32 8)
+ %res2 = add <4 x i64> %res, %res1
+ ret <4 x i64> %res2
}
declare <4 x float> @llvm.x86.avx512.gather3div4.sf(<4 x float>, i8*, <2 x i64>, i8, i32)
@@ -434,19 +434,19 @@ define <2 x double>@test_int_x86_avx512_
ret <2 x double> %res2
}
-declare <4 x i32> @llvm.x86.avx512.gather3siv2.di(<2 x i64>, i8*, <4 x i32>, i8, i32)
+declare <2 x i64> @llvm.x86.avx512.gather3siv2.di(<2 x i64>, i8*, <4 x i32>, i8, i32)
-define <4 x i32>@test_int_x86_avx512_gather3siv2_di(<2 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3) {
+define <2 x i64>@test_int_x86_avx512_gather3siv2_di(<2 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_gather3siv2_di:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovb %esi, %k1
; CHECK-NEXT: vpgatherdq (%rdi,%xmm1,8), %xmm0 {%k1}
-; CHECK-NEXT: vpaddd %xmm0, %xmm0, %xmm0
+; CHECK-NEXT: vpaddq %xmm0, %xmm0, %xmm0
; CHECK-NEXT: retq
- %res = call <4 x i32> @llvm.x86.avx512.gather3siv2.di(<2 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
- %res1 = call <4 x i32> @llvm.x86.avx512.gather3siv2.di(<2 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
- %res2 = add <4 x i32> %res, %res1
- ret <4 x i32> %res2
+ %res = call <2 x i64> @llvm.x86.avx512.gather3siv2.di(<2 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
+ %res1 = call <2 x i64> @llvm.x86.avx512.gather3siv2.di(<2 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
+ %res2 = add <2 x i64> %res, %res1
+ ret <2 x i64> %res2
}
declare <4 x double> @llvm.x86.avx512.gather3siv4.df(<4 x double>, i8*, <4 x i32>, i8, i32)
@@ -467,19 +467,19 @@ define <4 x double>@test_int_x86_avx512_
ret <4 x double> %res2
}
-declare <8 x i32> @llvm.x86.avx512.gather3siv4.di(<4 x i64>, i8*, <4 x i32>, i8, i32)
+declare <4 x i64> @llvm.x86.avx512.gather3siv4.di(<4 x i64>, i8*, <4 x i32>, i8, i32)
-define <8 x i32>@test_int_x86_avx512_gather3siv4_di(<4 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3) {
+define <4 x i64>@test_int_x86_avx512_gather3siv4_di(<4 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_gather3siv4_di:
; CHECK: ## BB#0:
; CHECK-NEXT: kmovb %esi, %k1
; CHECK-NEXT: vpgatherdq (%rdi,%xmm1,8), %ymm0 {%k1}
-; CHECK-NEXT: vpaddd %ymm0, %ymm0, %ymm0
+; CHECK-NEXT: vpaddq %ymm0, %ymm0, %ymm0
; CHECK-NEXT: retq
- %res = call <8 x i32> @llvm.x86.avx512.gather3siv4.di(<4 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
- %res1 = call <8 x i32> @llvm.x86.avx512.gather3siv4.di(<4 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
- %res2 = add <8 x i32> %res, %res1
- ret <8 x i32> %res2
+ %res = call <4 x i64> @llvm.x86.avx512.gather3siv4.di(<4 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
+ %res1 = call <4 x i64> @llvm.x86.avx512.gather3siv4.di(<4 x i64> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 8)
+ %res2 = add <4 x i64> %res, %res1
+ ret <4 x i64> %res2
}
declare <4 x float> @llvm.x86.avx512.gather3siv4.sf(<4 x float>, i8*, <4 x i32>, i8, i32)
More information about the llvm-commits
mailing list