| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py |
| ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX512F |
| ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=CHECK,AVX512DQ |
| ; 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) { |
| ; AVX512F-LABEL: PR32546: |
| ; AVX512F: ## %bb.0: ## %entry |
| ; AVX512F-NEXT: ## kill: def $ymm3 killed $ymm3 def $zmm3 |
| ; AVX512F-NEXT: ## kill: def $ymm2 killed $ymm2 def $zmm2 |
| ; AVX512F-NEXT: ## kill: def $ymm1 killed $ymm1 def $zmm1 |
| ; AVX512F-NEXT: ## kill: def $ymm0 killed $ymm0 def $zmm0 |
| ; AVX512F-NEXT: vcmpltps %zmm3, %zmm2, %k1 |
| ; AVX512F-NEXT: vcmpltps %zmm1, %zmm0, %k0 {%k1} |
| ; AVX512F-NEXT: kmovw %k0, %eax |
| ; AVX512F-NEXT: movzbl %al, %eax |
| ; AVX512F-NEXT: vmovd %eax, %xmm0 |
| ; AVX512F-NEXT: vpbroadcastd %xmm0, %ymm0 |
| ; AVX512F-NEXT: retq |
| ; |
| ; AVX512DQ-LABEL: PR32546: |
| ; AVX512DQ: ## %bb.0: ## %entry |
| ; AVX512DQ-NEXT: ## kill: def $ymm3 killed $ymm3 def $zmm3 |
| ; AVX512DQ-NEXT: ## kill: def $ymm2 killed $ymm2 def $zmm2 |
| ; AVX512DQ-NEXT: ## kill: def $ymm1 killed $ymm1 def $zmm1 |
| ; AVX512DQ-NEXT: ## kill: def $ymm0 killed $ymm0 def $zmm0 |
| ; AVX512DQ-NEXT: vcmpltps %zmm3, %zmm2, %k1 |
| ; AVX512DQ-NEXT: vcmpltps %zmm1, %zmm0, %k0 {%k1} |
| ; AVX512DQ-NEXT: kmovb %k0, %eax |
| ; AVX512DQ-NEXT: vmovd %eax, %xmm0 |
| ; AVX512DQ-NEXT: vpbroadcastd %xmm0, %ymm0 |
| ; AVX512DQ-NEXT: retq |
| ; |
| ; AVX512DQVL-LABEL: PR32546: |
| ; AVX512DQVL: ## %bb.0: ## %entry |
| ; AVX512DQVL-NEXT: vcmpltps %ymm3, %ymm2, %k1 |
| ; AVX512DQVL-NEXT: vcmpltps %ymm1, %ymm0, %k0 {%k1} |
| ; 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 |
| } |
| declare i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float>, <8 x float>, i32, i8) |