9512525947
When we use mask compare intrinsics under strict FP option, the masked elements shouldn't raise any exception. So, we cann't replace the intrinsic with a full compare + "and" operation. Reviewed By: craig.topper Differential Revision: https://reviews.llvm.org/D85385
139 lines
6.1 KiB
LLVM
139 lines
6.1 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512vl | FileCheck %s --check-prefixes=CHECK,AVX512VL
|
|
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f,+avx512vl,+avx512dq | FileCheck %s --check-prefixes=CHECK,AVX512DQVL
|
|
|
|
define <4 x i64> @PR32546(<8 x float> %a, <8 x float> %b, <8 x float> %c, <8 x float> %d) {
|
|
; AVX512VL-LABEL: PR32546:
|
|
; AVX512VL: ## %bb.0: ## %entry
|
|
; AVX512VL-NEXT: vcmpltps %ymm1, %ymm0, %k0
|
|
; AVX512VL-NEXT: vcmpltps %ymm3, %ymm2, %k1
|
|
; AVX512VL-NEXT: kandw %k0, %k1, %k0
|
|
; AVX512VL-NEXT: kmovw %k0, %eax
|
|
; AVX512VL-NEXT: movzbl %al, %eax
|
|
; AVX512VL-NEXT: vpbroadcastd %eax, %ymm0
|
|
; AVX512VL-NEXT: retq
|
|
;
|
|
; AVX512DQVL-LABEL: PR32546:
|
|
; AVX512DQVL: ## %bb.0: ## %entry
|
|
; AVX512DQVL-NEXT: vcmpltps %ymm1, %ymm0, %k0
|
|
; AVX512DQVL-NEXT: vcmpltps %ymm3, %ymm2, %k1
|
|
; AVX512DQVL-NEXT: kandb %k0, %k1, %k0
|
|
; AVX512DQVL-NEXT: kmovb %k0, %eax
|
|
; AVX512DQVL-NEXT: vpbroadcastd %eax, %ymm0
|
|
; AVX512DQVL-NEXT: retq
|
|
entry:
|
|
%0 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %a, <8 x float> %b, i32 1, i8 -1)
|
|
%1 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %c, <8 x float> %d, i32 1, i8 -1)
|
|
%and17 = and i8 %1, %0
|
|
%and = zext i8 %and17 to i32
|
|
%2 = insertelement <8 x i32> undef, i32 %and, i32 0
|
|
%vecinit7.i = shufflevector <8 x i32> %2, <8 x i32> undef, <8 x i32> zeroinitializer
|
|
%3 = bitcast <8 x i32> %vecinit7.i to <4 x i64>
|
|
ret <4 x i64> %3
|
|
}
|
|
|
|
define void @PR32547(<8 x float> %a, <8 x float> %b, <8 x float> %c, <8 x float> %d, float* %p) {
|
|
; CHECK-LABEL: PR32547:
|
|
; CHECK: ## %bb.0: ## %entry
|
|
; CHECK-NEXT: vcmpltps %ymm1, %ymm0, %k0
|
|
; CHECK-NEXT: vcmpltps %ymm3, %ymm2, %k1
|
|
; CHECK-NEXT: kunpckbw %k1, %k0, %k1
|
|
; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0
|
|
; CHECK-NEXT: vmovaps %zmm0, (%rdi) {%k1}
|
|
; CHECK-NEXT: vzeroupper
|
|
; CHECK-NEXT: retq
|
|
entry:
|
|
%0 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %a, <8 x float> %b, i32 1, i8 -1)
|
|
%1 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %c, <8 x float> %d, i32 1, i8 -1)
|
|
%conv.i = zext i8 %0 to i16
|
|
%conv.i18 = zext i8 %1 to i16
|
|
%shl = shl nuw i16 %conv.i, 8
|
|
%or = or i16 %shl, %conv.i18
|
|
%2 = bitcast float* %p to <16 x float>*
|
|
%3 = bitcast i16 %or to <16 x i1>
|
|
tail call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> zeroinitializer, <16 x float>* %2, i32 64, <16 x i1> %3)
|
|
ret void
|
|
}
|
|
|
|
define void @PR32547_swap(<8 x float> %a, <8 x float> %b, <8 x float> %c, <8 x float> %d, float* %p) {
|
|
; CHECK-LABEL: PR32547_swap:
|
|
; CHECK: ## %bb.0: ## %entry
|
|
; CHECK-NEXT: vcmpltps %ymm1, %ymm0, %k0
|
|
; CHECK-NEXT: vcmpltps %ymm3, %ymm2, %k1
|
|
; CHECK-NEXT: kunpckbw %k1, %k0, %k1
|
|
; CHECK-NEXT: vxorps %xmm0, %xmm0, %xmm0
|
|
; CHECK-NEXT: vmovaps %zmm0, (%rdi) {%k1}
|
|
; CHECK-NEXT: vzeroupper
|
|
; CHECK-NEXT: retq
|
|
entry:
|
|
%0 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %a, <8 x float> %b, i32 1, i8 -1)
|
|
%1 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %c, <8 x float> %d, i32 1, i8 -1)
|
|
%conv.i = zext i8 %0 to i16
|
|
%conv.i18 = zext i8 %1 to i16
|
|
%shl = shl nuw i16 %conv.i, 8
|
|
%or = or i16 %conv.i18, %shl
|
|
%2 = bitcast float* %p to <16 x float>*
|
|
%3 = bitcast i16 %or to <16 x i1>
|
|
tail call void @llvm.masked.store.v16f32.p0v16f32(<16 x float> zeroinitializer, <16 x float>* %2, i32 64, <16 x i1> %3)
|
|
ret void
|
|
}
|
|
|
|
define void @mask_cmp_128(<4 x float> %a, <4 x float> %b, <4 x float> %c, <4 x float> %d, float* %p) {
|
|
; AVX512VL-LABEL: mask_cmp_128:
|
|
; AVX512VL: ## %bb.0: ## %entry
|
|
; AVX512VL-NEXT: vcmpltps %xmm1, %xmm0, %k0
|
|
; AVX512VL-NEXT: kmovw %k0, %eax
|
|
; AVX512VL-NEXT: vcmpltps %xmm3, %xmm2, %k0
|
|
; AVX512VL-NEXT: shlb $4, %al
|
|
; AVX512VL-NEXT: kmovw %eax, %k1
|
|
; AVX512VL-NEXT: korw %k1, %k0, %k1
|
|
; AVX512VL-NEXT: vxorps %xmm0, %xmm0, %xmm0
|
|
; AVX512VL-NEXT: vmovaps %ymm0, (%rdi) {%k1}
|
|
; AVX512VL-NEXT: vzeroupper
|
|
; AVX512VL-NEXT: retq
|
|
;
|
|
; AVX512DQVL-LABEL: mask_cmp_128:
|
|
; AVX512DQVL: ## %bb.0: ## %entry
|
|
; AVX512DQVL-NEXT: vcmpltps %xmm1, %xmm0, %k0
|
|
; AVX512DQVL-NEXT: vcmpltps %xmm3, %xmm2, %k1
|
|
; AVX512DQVL-NEXT: kshiftlb $4, %k0, %k0
|
|
; AVX512DQVL-NEXT: korb %k0, %k1, %k1
|
|
; AVX512DQVL-NEXT: vxorps %xmm0, %xmm0, %xmm0
|
|
; AVX512DQVL-NEXT: vmovaps %ymm0, (%rdi) {%k1}
|
|
; AVX512DQVL-NEXT: vzeroupper
|
|
; AVX512DQVL-NEXT: retq
|
|
entry:
|
|
%0 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float> %a, <4 x float> %b, i32 1, i8 -1)
|
|
%1 = tail call i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float> %c, <4 x float> %d, i32 1, i8 -1)
|
|
%shl = shl nuw i8 %0, 4
|
|
%or = or i8 %1, %shl
|
|
%2 = bitcast float* %p to <8 x float>*
|
|
%3 = bitcast i8 %or to <8 x i1>
|
|
tail call void @llvm.masked.store.v8f32.p0v8f32(<8 x float> zeroinitializer, <8 x float>* %2, i32 64, <8 x i1> %3)
|
|
ret void
|
|
}
|
|
|
|
define <16 x float> @mask_cmp_512(<16 x float> %a, <16 x float> %b, <16 x float> %c, <16 x float> %d, float* %p) {
|
|
; CHECK-LABEL: mask_cmp_512:
|
|
; CHECK: ## %bb.0: ## %entry
|
|
; CHECK-NEXT: vcmpltps {sae}, %zmm1, %zmm0, %k0
|
|
; CHECK-NEXT: vcmpltps %zmm3, %zmm2, %k1
|
|
; CHECK-NEXT: kxnorw %k1, %k0, %k1
|
|
; CHECK-NEXT: vmovaps (%rdi), %zmm0 {%k1} {z}
|
|
; CHECK-NEXT: retq
|
|
entry:
|
|
%0 = tail call i16 @llvm.x86.avx512.mask.cmp.ps.512(<16 x float> %a, <16 x float> %b, i32 1, i16 -1, i32 8)
|
|
%1 = tail call i16 @llvm.x86.avx512.mask.cmp.ps.512(<16 x float> %c, <16 x float> %d, i32 1, i16 -1, i32 4)
|
|
%2 = bitcast float* %p to <16 x float>*
|
|
%3 = load <16 x float>, <16 x float>* %2
|
|
%4 = xor i16 %0, %1
|
|
%5 = bitcast i16 %4 to <16 x i1>
|
|
%6 = select <16 x i1> %5, <16 x float> zeroinitializer, <16 x float> %3
|
|
ret <16 x float> %6
|
|
}
|
|
declare i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float>, <4 x float>, i32, i8)
|
|
declare i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float>, <8 x float>, i32, i8)
|
|
declare i16 @llvm.x86.avx512.mask.cmp.ps.512(<16 x float>, <16 x float>, i32, i16, i32)
|
|
declare void @llvm.masked.store.v8f32.p0v8f32(<8 x float>, <8 x float>*, i32, <8 x i1>)
|
|
declare void @llvm.masked.store.v16f32.p0v16f32(<16 x float>, <16 x float>*, i32, <16 x i1>)
|