diff options
author | Craig Topper <craig.topper@gmail.com> | 2016-05-22 00:39:33 +0000 |
---|---|---|
committer | Craig Topper <craig.topper@gmail.com> | 2016-05-22 00:39:33 +0000 |
commit | 9c5f275934a2a28181832d12d3578b19e52fec0c (patch) | |
tree | 2fe2b4c6c8f93e9f040f7a6912bde3ec49d8d91c /test/CodeGen/X86 | |
parent | 6a96e9fdad5ff57dd32848e7331bace12e861250 (diff) |
[X86] Add a common check-prefix to both run lines on a test so identical checks appear just once.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@270345 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/X86')
-rw-r--r-- | test/CodeGen/X86/avx512-vec-cmp.ll | 529 |
1 files changed, 168 insertions, 361 deletions
diff --git a/test/CodeGen/X86/avx512-vec-cmp.ll b/test/CodeGen/X86/avx512-vec-cmp.ll index 73a8f26fbca..8f29c9d5c16 100644 --- a/test/CodeGen/X86/avx512-vec-cmp.ll +++ b/test/CodeGen/X86/avx512-vec-cmp.ll @@ -1,59 +1,38 @@ ; NOTE: Assertions have been autogenerated by update_llc_test_checks.py -; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s --check-prefix=KNL -; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=skx | FileCheck %s --check-prefix=SKX +; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s --check-prefix=CHECK --check-prefix=KNL +; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=skx | FileCheck %s --check-prefix=CHECK --check-prefix=SKX define <16 x float> @test1(<16 x float> %x, <16 x float> %y) nounwind { -; KNL-LABEL: test1: -; KNL: ## BB#0: -; KNL-NEXT: vcmpleps %zmm1, %zmm0, %k1 -; KNL-NEXT: vmovaps %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test1: -; SKX: ## BB#0: -; SKX-NEXT: vcmpleps %zmm1, %zmm0, %k1 -; SKX-NEXT: vmovaps %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test1: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmpleps %zmm1, %zmm0, %k1 +; CHECK-NEXT: vmovaps %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask = fcmp ole <16 x float> %x, %y %max = select <16 x i1> %mask, <16 x float> %x, <16 x float> %y ret <16 x float> %max } define <8 x double> @test2(<8 x double> %x, <8 x double> %y) nounwind { -; KNL-LABEL: test2: -; KNL: ## BB#0: -; KNL-NEXT: vcmplepd %zmm1, %zmm0, %k1 -; KNL-NEXT: vmovapd %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test2: -; SKX: ## BB#0: -; SKX-NEXT: vcmplepd %zmm1, %zmm0, %k1 -; SKX-NEXT: vmovapd %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test2: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmplepd %zmm1, %zmm0, %k1 +; CHECK-NEXT: vmovapd %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask = fcmp ole <8 x double> %x, %y %max = select <8 x i1> %mask, <8 x double> %x, <8 x double> %y ret <8 x double> %max } define <16 x i32> @test3(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %yp) nounwind { -; KNL-LABEL: test3: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpeqd (%rdi), %zmm0, %k1 -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test3: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpeqd (%rdi), %zmm0, %k1 -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test3: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpeqd (%rdi), %zmm0, %k1 +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %y = load <16 x i32>, <16 x i32>* %yp, align 4 %mask = icmp eq <16 x i32> %x, %y %max = select <16 x i1> %mask, <16 x i32> %x, <16 x i32> %x1 @@ -61,57 +40,36 @@ define <16 x i32> @test3(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %yp) nounwin } define <16 x i32> @test4_unsigned(<16 x i32> %x, <16 x i32> %y, <16 x i32> %x1) nounwind { -; KNL-LABEL: test4_unsigned: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpnltud %zmm1, %zmm0, %k1 -; KNL-NEXT: vmovdqa32 %zmm2, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test4_unsigned: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpnltud %zmm1, %zmm0, %k1 -; SKX-NEXT: vmovdqa32 %zmm2, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test4_unsigned: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpnltud %zmm1, %zmm0, %k1 +; CHECK-NEXT: vmovdqa32 %zmm2, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask = icmp uge <16 x i32> %x, %y %max = select <16 x i1> %mask, <16 x i32> %x1, <16 x i32> %y ret <16 x i32> %max } define <8 x i64> @test5(<8 x i64> %x, <8 x i64> %y) nounwind { -; KNL-LABEL: test5: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpeqq %zmm1, %zmm0, %k1 -; KNL-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test5: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpeqq %zmm1, %zmm0, %k1 -; SKX-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test5: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpeqq %zmm1, %zmm0, %k1 +; CHECK-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask = icmp eq <8 x i64> %x, %y %max = select <8 x i1> %mask, <8 x i64> %x, <8 x i64> %y ret <8 x i64> %max } define <8 x i64> @test6_unsigned(<8 x i64> %x, <8 x i64> %y, <8 x i64> %x1) nounwind { -; KNL-LABEL: test6_unsigned: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpnleuq %zmm1, %zmm0, %k1 -; KNL-NEXT: vmovdqa64 %zmm2, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test6_unsigned: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpnleuq %zmm1, %zmm0, %k1 -; SKX-NEXT: vmovdqa64 %zmm2, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test6_unsigned: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpnleuq %zmm1, %zmm0, %k1 +; CHECK-NEXT: vmovdqa64 %zmm2, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask = icmp ugt <8 x i64> %x, %y %max = select <8 x i1> %mask, <8 x i64> %x1, <8 x i64> %y ret <8 x i64> %max @@ -196,36 +154,23 @@ define <8 x float> @test10(<8 x float> %x, <8 x float> %y) nounwind { } define <8 x i32> @test11_unsigned(<8 x i32> %x, <8 x i32> %y) nounwind { -; KNL-LABEL: test11_unsigned: -; KNL: ## BB#0: -; KNL-NEXT: vpmaxud %ymm1, %ymm0, %ymm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test11_unsigned: -; SKX: ## BB#0: -; SKX-NEXT: vpmaxud %ymm1, %ymm0, %ymm0 -; SKX-NEXT: retq +; CHECK-LABEL: test11_unsigned: +; CHECK: ## BB#0: +; CHECK-NEXT: vpmaxud %ymm1, %ymm0, %ymm0 +; CHECK-NEXT: retq %mask = icmp ugt <8 x i32> %x, %y %max = select <8 x i1> %mask, <8 x i32> %x, <8 x i32> %y ret <8 x i32> %max } define i16 @test12(<16 x i64> %a, <16 x i64> %b) nounwind { -; KNL-LABEL: test12: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpeqq %zmm2, %zmm0, %k0 -; KNL-NEXT: vpcmpeqq %zmm3, %zmm1, %k1 -; KNL-NEXT: kunpckbw %k0, %k1, %k0 -; KNL-NEXT: kmovw %k0, %eax -; KNL-NEXT: retq -; -; SKX-LABEL: test12: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpeqq %zmm2, %zmm0, %k0 -; SKX-NEXT: vpcmpeqq %zmm3, %zmm1, %k1 -; SKX-NEXT: kunpckbw %k0, %k1, %k0 -; SKX-NEXT: kmovw %k0, %eax -; SKX-NEXT: retq +; CHECK-LABEL: test12: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpeqq %zmm2, %zmm0, %k0 +; CHECK-NEXT: vpcmpeqq %zmm3, %zmm1, %k1 +; CHECK-NEXT: kunpckbw %k0, %k1, %k0 +; CHECK-NEXT: kmovw %k0, %eax +; CHECK-NEXT: retq %res = icmp eq <16 x i64> %a, %b %res1 = bitcast <16 x i1> %res to i16 ret i16 %res1 @@ -258,17 +203,11 @@ define i64 @test12_v64i16(<64 x i16> %a, <64 x i16> %b) nounwind { } define <16 x i32> @test13(<16 x float>%a, <16 x float>%b) -; KNL-LABEL: test13: -; KNL: ## BB#0: -; KNL-NEXT: vcmpeqps %zmm1, %zmm0, %k1 -; KNL-NEXT: vpbroadcastd {{.*}}(%rip), %zmm0 {%k1} {z} -; KNL-NEXT: retq -; -; SKX-LABEL: test13: -; SKX: ## BB#0: -; SKX-NEXT: vcmpeqps %zmm1, %zmm0, %k1 -; SKX-NEXT: vpbroadcastd {{.*}}(%rip), %zmm0 {%k1} {z} -; SKX-NEXT: retq +; CHECK-LABEL: test13: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmpeqps %zmm1, %zmm0, %k1 +; CHECK-NEXT: vpbroadcastd {{.*}}(%rip), %zmm0 {%k1} {z} +; CHECK-NEXT: retq { %cmpvector_i = fcmp oeq <16 x float> %a, %b %conv = zext <16 x i1> %cmpvector_i to <16 x i32> @@ -276,19 +215,12 @@ define <16 x i32> @test13(<16 x float>%a, <16 x float>%b) } define <16 x i32> @test14(<16 x i32>%a, <16 x i32>%b) { -; KNL-LABEL: test14: -; KNL: ## BB#0: -; KNL-NEXT: vpsubd %zmm1, %zmm0, %zmm1 -; KNL-NEXT: vpcmpgtd %zmm0, %zmm1, %k1 -; KNL-NEXT: vmovdqa32 %zmm1, %zmm0 {%k1} {z} -; KNL-NEXT: retq -; -; SKX-LABEL: test14: -; SKX: ## BB#0: -; SKX-NEXT: vpsubd %zmm1, %zmm0, %zmm1 -; SKX-NEXT: vpcmpgtd %zmm0, %zmm1, %k1 -; SKX-NEXT: vmovdqa32 %zmm1, %zmm0 {%k1} {z} -; SKX-NEXT: retq +; CHECK-LABEL: test14: +; CHECK: ## BB#0: +; CHECK-NEXT: vpsubd %zmm1, %zmm0, %zmm1 +; CHECK-NEXT: vpcmpgtd %zmm0, %zmm1, %k1 +; CHECK-NEXT: vmovdqa32 %zmm1, %zmm0 {%k1} {z} +; CHECK-NEXT: retq %sub_r = sub <16 x i32> %a, %b %cmp.i2.i = icmp sgt <16 x i32> %sub_r, %a %sext.i3.i = sext <16 x i1> %cmp.i2.i to <16 x i32> @@ -298,19 +230,12 @@ define <16 x i32> @test14(<16 x i32>%a, <16 x i32>%b) { } define <8 x i64> @test15(<8 x i64>%a, <8 x i64>%b) { -; KNL-LABEL: test15: -; KNL: ## BB#0: -; KNL-NEXT: vpsubq %zmm1, %zmm0, %zmm1 -; KNL-NEXT: vpcmpgtq %zmm0, %zmm1, %k1 -; KNL-NEXT: vmovdqa64 %zmm1, %zmm0 {%k1} {z} -; KNL-NEXT: retq -; -; SKX-LABEL: test15: -; SKX: ## BB#0: -; SKX-NEXT: vpsubq %zmm1, %zmm0, %zmm1 -; SKX-NEXT: vpcmpgtq %zmm0, %zmm1, %k1 -; SKX-NEXT: vmovdqa64 %zmm1, %zmm0 {%k1} {z} -; SKX-NEXT: retq +; CHECK-LABEL: test15: +; CHECK: ## BB#0: +; CHECK-NEXT: vpsubq %zmm1, %zmm0, %zmm1 +; CHECK-NEXT: vpcmpgtq %zmm0, %zmm1, %k1 +; CHECK-NEXT: vmovdqa64 %zmm1, %zmm0 {%k1} {z} +; CHECK-NEXT: retq %sub_r = sub <8 x i64> %a, %b %cmp.i2.i = icmp sgt <8 x i64> %sub_r, %a %sext.i3.i = sext <8 x i1> %cmp.i2.i to <8 x i64> @@ -320,38 +245,24 @@ define <8 x i64> @test15(<8 x i64>%a, <8 x i64>%b) { } define <16 x i32> @test16(<16 x i32> %x, <16 x i32> %y, <16 x i32> %x1) nounwind { -; KNL-LABEL: test16: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpled %zmm0, %zmm1, %k1 -; KNL-NEXT: vmovdqa32 %zmm2, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test16: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpled %zmm0, %zmm1, %k1 -; SKX-NEXT: vmovdqa32 %zmm2, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test16: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpled %zmm0, %zmm1, %k1 +; CHECK-NEXT: vmovdqa32 %zmm2, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask = icmp sge <16 x i32> %x, %y %max = select <16 x i1> %mask, <16 x i32> %x1, <16 x i32> %y ret <16 x i32> %max } define <16 x i32> @test17(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %y.ptr) nounwind { -; KNL-LABEL: test17: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpgtd (%rdi), %zmm0, %k1 -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test17: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpgtd (%rdi), %zmm0, %k1 -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test17: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpgtd (%rdi), %zmm0, %k1 +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %y = load <16 x i32>, <16 x i32>* %y.ptr, align 4 %mask = icmp sgt <16 x i32> %x, %y %max = select <16 x i1> %mask, <16 x i32> %x, <16 x i32> %x1 @@ -359,19 +270,12 @@ define <16 x i32> @test17(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %y.ptr) nou } define <16 x i32> @test18(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %y.ptr) nounwind { -; KNL-LABEL: test18: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpled (%rdi), %zmm0, %k1 -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test18: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpled (%rdi), %zmm0, %k1 -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test18: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpled (%rdi), %zmm0, %k1 +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %y = load <16 x i32>, <16 x i32>* %y.ptr, align 4 %mask = icmp sle <16 x i32> %x, %y %max = select <16 x i1> %mask, <16 x i32> %x, <16 x i32> %x1 @@ -379,19 +283,12 @@ define <16 x i32> @test18(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %y.ptr) nou } define <16 x i32> @test19(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %y.ptr) nounwind { -; KNL-LABEL: test19: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpleud (%rdi), %zmm0, %k1 -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test19: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpleud (%rdi), %zmm0, %k1 -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test19: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpleud (%rdi), %zmm0, %k1 +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %y = load <16 x i32>, <16 x i32>* %y.ptr, align 4 %mask = icmp ule <16 x i32> %x, %y %max = select <16 x i1> %mask, <16 x i32> %x, <16 x i32> %x1 @@ -399,21 +296,13 @@ define <16 x i32> @test19(<16 x i32> %x, <16 x i32> %x1, <16 x i32>* %y.ptr) nou } define <16 x i32> @test20(<16 x i32> %x, <16 x i32> %y, <16 x i32> %x1, <16 x i32> %y1) nounwind { -; KNL-LABEL: test20: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpeqd %zmm1, %zmm0, %k1 -; KNL-NEXT: vpcmpeqd %zmm3, %zmm2, %k1 {%k1} -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test20: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpeqd %zmm1, %zmm0, %k1 -; SKX-NEXT: vpcmpeqd %zmm3, %zmm2, %k1 {%k1} -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test20: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpeqd %zmm1, %zmm0, %k1 +; CHECK-NEXT: vpcmpeqd %zmm3, %zmm2, %k1 {%k1} +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask1 = icmp eq <16 x i32> %x1, %y1 %mask0 = icmp eq <16 x i32> %x, %y %mask = select <16 x i1> %mask0, <16 x i1> %mask1, <16 x i1> zeroinitializer @@ -422,21 +311,13 @@ define <16 x i32> @test20(<16 x i32> %x, <16 x i32> %y, <16 x i32> %x1, <16 x i3 } define <8 x i64> @test21(<8 x i64> %x, <8 x i64> %y, <8 x i64> %x1, <8 x i64> %y1) nounwind { -; KNL-LABEL: test21: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpleq %zmm1, %zmm0, %k1 -; KNL-NEXT: vpcmpleq %zmm2, %zmm3, %k1 {%k1} -; KNL-NEXT: vmovdqa64 %zmm0, %zmm2 {%k1} -; KNL-NEXT: vmovaps %zmm2, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test21: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpleq %zmm1, %zmm0, %k1 -; SKX-NEXT: vpcmpleq %zmm2, %zmm3, %k1 {%k1} -; SKX-NEXT: vmovdqa64 %zmm0, %zmm2 {%k1} -; SKX-NEXT: vmovaps %zmm2, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test21: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpleq %zmm1, %zmm0, %k1 +; CHECK-NEXT: vpcmpleq %zmm2, %zmm3, %k1 {%k1} +; CHECK-NEXT: vmovdqa64 %zmm0, %zmm2 {%k1} +; CHECK-NEXT: vmovaps %zmm2, %zmm0 +; CHECK-NEXT: retq %mask1 = icmp sge <8 x i64> %x1, %y1 %mask0 = icmp sle <8 x i64> %x, %y %mask = select <8 x i1> %mask0, <8 x i1> %mask1, <8 x i1> zeroinitializer @@ -445,21 +326,13 @@ define <8 x i64> @test21(<8 x i64> %x, <8 x i64> %y, <8 x i64> %x1, <8 x i64> %y } define <8 x i64> @test22(<8 x i64> %x, <8 x i64>* %y.ptr, <8 x i64> %x1, <8 x i64> %y1) nounwind { -; KNL-LABEL: test22: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpgtq %zmm2, %zmm1, %k1 -; KNL-NEXT: vpcmpgtq (%rdi), %zmm0, %k1 {%k1} -; KNL-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test22: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpgtq %zmm2, %zmm1, %k1 -; SKX-NEXT: vpcmpgtq (%rdi), %zmm0, %k1 {%k1} -; SKX-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test22: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpgtq %zmm2, %zmm1, %k1 +; CHECK-NEXT: vpcmpgtq (%rdi), %zmm0, %k1 {%k1} +; CHECK-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask1 = icmp sgt <8 x i64> %x1, %y1 %y = load <8 x i64>, <8 x i64>* %y.ptr, align 4 %mask0 = icmp sgt <8 x i64> %x, %y @@ -469,21 +342,13 @@ define <8 x i64> @test22(<8 x i64> %x, <8 x i64>* %y.ptr, <8 x i64> %x1, <8 x i6 } define <16 x i32> @test23(<16 x i32> %x, <16 x i32>* %y.ptr, <16 x i32> %x1, <16 x i32> %y1) nounwind { -; KNL-LABEL: test23: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpled %zmm1, %zmm2, %k1 -; KNL-NEXT: vpcmpleud (%rdi), %zmm0, %k1 {%k1} -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test23: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpled %zmm1, %zmm2, %k1 -; SKX-NEXT: vpcmpleud (%rdi), %zmm0, %k1 {%k1} -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test23: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpled %zmm1, %zmm2, %k1 +; CHECK-NEXT: vpcmpleud (%rdi), %zmm0, %k1 {%k1} +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask1 = icmp sge <16 x i32> %x1, %y1 %y = load <16 x i32>, <16 x i32>* %y.ptr, align 4 %mask0 = icmp ule <16 x i32> %x, %y @@ -493,19 +358,12 @@ define <16 x i32> @test23(<16 x i32> %x, <16 x i32>* %y.ptr, <16 x i32> %x1, <16 } define <8 x i64> @test24(<8 x i64> %x, <8 x i64> %x1, i64* %yb.ptr) nounwind { -; KNL-LABEL: test24: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpeqq (%rdi){1to8}, %zmm0, %k1 -; KNL-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test24: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpeqq (%rdi){1to8}, %zmm0, %k1 -; SKX-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test24: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpeqq (%rdi){1to8}, %zmm0, %k1 +; CHECK-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %yb = load i64, i64* %yb.ptr, align 4 %y.0 = insertelement <8 x i64> undef, i64 %yb, i32 0 %y = shufflevector <8 x i64> %y.0, <8 x i64> undef, <8 x i32> zeroinitializer @@ -515,19 +373,12 @@ define <8 x i64> @test24(<8 x i64> %x, <8 x i64> %x1, i64* %yb.ptr) nounwind { } define <16 x i32> @test25(<16 x i32> %x, i32* %yb.ptr, <16 x i32> %x1) nounwind { -; KNL-LABEL: test25: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpled (%rdi){1to16}, %zmm0, %k1 -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test25: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpled (%rdi){1to16}, %zmm0, %k1 -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test25: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpled (%rdi){1to16}, %zmm0, %k1 +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %yb = load i32, i32* %yb.ptr, align 4 %y.0 = insertelement <16 x i32> undef, i32 %yb, i32 0 %y = shufflevector <16 x i32> %y.0, <16 x i32> undef, <16 x i32> zeroinitializer @@ -537,21 +388,13 @@ define <16 x i32> @test25(<16 x i32> %x, i32* %yb.ptr, <16 x i32> %x1) nounwind } define <16 x i32> @test26(<16 x i32> %x, i32* %yb.ptr, <16 x i32> %x1, <16 x i32> %y1) nounwind { -; KNL-LABEL: test26: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpled %zmm1, %zmm2, %k1 -; KNL-NEXT: vpcmpgtd (%rdi){1to16}, %zmm0, %k1 {%k1} -; KNL-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test26: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpled %zmm1, %zmm2, %k1 -; SKX-NEXT: vpcmpgtd (%rdi){1to16}, %zmm0, %k1 {%k1} -; SKX-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test26: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpled %zmm1, %zmm2, %k1 +; CHECK-NEXT: vpcmpgtd (%rdi){1to16}, %zmm0, %k1 {%k1} +; CHECK-NEXT: vmovdqa32 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask1 = icmp sge <16 x i32> %x1, %y1 %yb = load i32, i32* %yb.ptr, align 4 %y.0 = insertelement <16 x i32> undef, i32 %yb, i32 0 @@ -563,21 +406,13 @@ define <16 x i32> @test26(<16 x i32> %x, i32* %yb.ptr, <16 x i32> %x1, <16 x i32 } define <8 x i64> @test27(<8 x i64> %x, i64* %yb.ptr, <8 x i64> %x1, <8 x i64> %y1) nounwind { -; KNL-LABEL: test27: -; KNL: ## BB#0: -; KNL-NEXT: vpcmpleq %zmm1, %zmm2, %k1 -; KNL-NEXT: vpcmpleq (%rdi){1to8}, %zmm0, %k1 {%k1} -; KNL-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test27: -; SKX: ## BB#0: -; SKX-NEXT: vpcmpleq %zmm1, %zmm2, %k1 -; SKX-NEXT: vpcmpleq (%rdi){1to8}, %zmm0, %k1 {%k1} -; SKX-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test27: +; CHECK: ## BB#0: +; CHECK-NEXT: vpcmpleq %zmm1, %zmm2, %k1 +; CHECK-NEXT: vpcmpleq (%rdi){1to8}, %zmm0, %k1 {%k1} +; CHECK-NEXT: vmovdqa64 %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %mask1 = icmp sge <8 x i64> %x1, %y1 %yb = load i64, i64* %yb.ptr, align 4 %y.0 = insertelement <8 x i64> undef, i64 %yb, i32 0 @@ -696,19 +531,12 @@ define <4 x double> @test32(<4 x double> %x, <4 x double> %x1, <4 x double>* %yp } define <8 x double> @test33(<8 x double> %x, <8 x double> %x1, <8 x double>* %yp) nounwind { -; KNL-LABEL: test33: -; KNL: ## BB#0: -; KNL-NEXT: vcmpltpd (%rdi), %zmm0, %k1 -; KNL-NEXT: vmovapd %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test33: -; SKX: ## BB#0: -; SKX-NEXT: vcmpltpd (%rdi), %zmm0, %k1 -; SKX-NEXT: vmovapd %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test33: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmpltpd (%rdi), %zmm0, %k1 +; CHECK-NEXT: vmovapd %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %y = load <8 x double>, <8 x double>* %yp, align 4 %mask = fcmp olt <8 x double> %x, %y %max = select <8 x i1> %mask, <8 x double> %x, <8 x double> %x1 @@ -756,19 +584,12 @@ define <8 x float> @test35(<8 x float> %x, <8 x float> %x1, <8 x float>* %yp) no } define <16 x float> @test36(<16 x float> %x, <16 x float> %x1, <16 x float>* %yp) nounwind { -; KNL-LABEL: test36: -; KNL: ## BB#0: -; KNL-NEXT: vcmpltps (%rdi), %zmm0, %k1 -; KNL-NEXT: vmovaps %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test36: -; SKX: ## BB#0: -; SKX-NEXT: vcmpltps (%rdi), %zmm0, %k1 -; SKX-NEXT: vmovaps %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test36: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmpltps (%rdi), %zmm0, %k1 +; CHECK-NEXT: vmovaps %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %y = load <16 x float>, <16 x float>* %yp, align 4 %mask = fcmp olt <16 x float> %x, %y %max = select <16 x i1> %mask, <16 x float> %x, <16 x float> %x1 @@ -776,19 +597,12 @@ define <16 x float> @test36(<16 x float> %x, <16 x float> %x1, <16 x float>* %yp } define <8 x double> @test37(<8 x double> %x, <8 x double> %x1, double* %ptr) nounwind { -; KNL-LABEL: test37: -; KNL: ## BB#0: -; KNL-NEXT: vcmpltpd (%rdi){1to8}, %zmm0, %k1 -; KNL-NEXT: vmovapd %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test37: -; SKX: ## BB#0: -; SKX-NEXT: vcmpltpd (%rdi){1to8}, %zmm0, %k1 -; SKX-NEXT: vmovapd %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test37: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmpltpd (%rdi){1to8}, %zmm0, %k1 +; CHECK-NEXT: vmovapd %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %a = load double, double* %ptr %v = insertelement <8 x double> undef, double %a, i32 0 @@ -849,19 +663,12 @@ define <2 x double> @test39(<2 x double> %x, <2 x double> %x1, double* %ptr) nou define <16 x float> @test40(<16 x float> %x, <16 x float> %x1, float* %ptr) nounwind { -; KNL-LABEL: test40: -; KNL: ## BB#0: -; KNL-NEXT: vcmpltps (%rdi){1to16}, %zmm0, %k1 -; KNL-NEXT: vmovaps %zmm0, %zmm1 {%k1} -; KNL-NEXT: vmovaps %zmm1, %zmm0 -; KNL-NEXT: retq -; -; SKX-LABEL: test40: -; SKX: ## BB#0: -; SKX-NEXT: vcmpltps (%rdi){1to16}, %zmm0, %k1 -; SKX-NEXT: vmovaps %zmm0, %zmm1 {%k1} -; SKX-NEXT: vmovaps %zmm1, %zmm0 -; SKX-NEXT: retq +; CHECK-LABEL: test40: +; CHECK: ## BB#0: +; CHECK-NEXT: vcmpltps (%rdi){1to16}, %zmm0, %k1 +; CHECK-NEXT: vmovaps %zmm0, %zmm1 {%k1} +; CHECK-NEXT: vmovaps %zmm1, %zmm0 +; CHECK-NEXT: retq %a = load float, float* %ptr %v = insertelement <16 x float> undef, float %a, i32 0 |