[llvm] CodeGen/X86: add missing mcpu to tests (PR #111865)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Oct 10 09:25:37 PDT 2024
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-x86
Author: Ramkumar Ramachandra (artagnon)
<details>
<summary>Changes</summary>
In order to faciliate working on core CodeGen infrastructure that improves the schedule of instructions in the default case where there is no scheduling information available in the machine model, add the missing -mcpu to existing tests, so that future patches doesn't unnecessarily change a huge number of tests. Note that the -mcpu argument is already present on all tests under most targets, including ARM and AArch64 targets.
Care has been taken to pick the right mcpu for each test, after studying its intent, and the patch should only affect the schedule of instructions in each test.
---
Patch is 799.45 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/111865.diff
35 Files Affected:
- (modified) llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll (+62-56)
- (modified) llvm/test/CodeGen/X86/bitcast-and-setcc-256.ll (+6-6)
- (modified) llvm/test/CodeGen/X86/div-rem-pair-recomposition-signed.ll (+372-354)
- (modified) llvm/test/CodeGen/X86/div-rem-pair-recomposition-unsigned.ll (+301-282)
- (modified) llvm/test/CodeGen/X86/fold-tied-op.ll (+62-57)
- (modified) llvm/test/CodeGen/X86/horizontal-sum.ll (+114-107)
- (modified) llvm/test/CodeGen/X86/is_fpclass.ll (+13-18)
- (modified) llvm/test/CodeGen/X86/lea-opt-cse4.ll (+18-10)
- (modified) llvm/test/CodeGen/X86/machine-cp.ll (+23-10)
- (modified) llvm/test/CodeGen/X86/madd.ll (+52-49)
- (modified) llvm/test/CodeGen/X86/masked_gather_scatter.ll (+206-314)
- (modified) llvm/test/CodeGen/X86/memcmp-more-load-pairs-x32.ll (+97-91)
- (modified) llvm/test/CodeGen/X86/memcmp-more-load-pairs.ll (+257-257)
- (modified) llvm/test/CodeGen/X86/memcmp.ll (+119-119)
- (modified) llvm/test/CodeGen/X86/midpoint-int-vec-256.ll (+160-160)
- (modified) llvm/test/CodeGen/X86/mul-constant-result.ll (+2-2)
- (modified) llvm/test/CodeGen/X86/mul-i512.ll (+568-565)
- (modified) llvm/test/CodeGen/X86/mul64.ll (+2-2)
- (modified) llvm/test/CodeGen/X86/pr62653.ll (+1-1)
- (modified) llvm/test/CodeGen/X86/rotate-multi.ll (+1-1)
- (modified) llvm/test/CodeGen/X86/sad.ll (+15-15)
- (modified) llvm/test/CodeGen/X86/sext-vsetcc.ll (+11-11)
- (modified) llvm/test/CodeGen/X86/smul_fix.ll (+54-56)
- (modified) llvm/test/CodeGen/X86/statepoint-live-in.ll (+1-2)
- (modified) llvm/test/CodeGen/X86/statepoint-regs.ll (+1-2)
- (modified) llvm/test/CodeGen/X86/ucmp.ll (+59-63)
- (modified) llvm/test/CodeGen/X86/umul-with-overflow.ll (+35-36)
- (modified) llvm/test/CodeGen/X86/umul_fix.ll (+75-78)
- (modified) llvm/test/CodeGen/X86/umulo-128-legalisation-lowering.ll (+27-25)
- (modified) llvm/test/CodeGen/X86/v8i1-masks.ll (+104-104)
- (modified) llvm/test/CodeGen/X86/vector-interleaved-store-i8-stride-5.ll (+835-835)
- (modified) llvm/test/CodeGen/X86/vector-reduce-or-cmp.ll (+7-7)
- (modified) llvm/test/CodeGen/X86/win-smallparams.ll (+4-4)
- (modified) llvm/test/CodeGen/X86/x86-interleaved-access.ll (+42-42)
- (modified) llvm/test/CodeGen/X86/xmulo.ll (+161-176)
``````````diff
diff --git a/llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll b/llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll
index 51ffeca52a6652..fe2e917434a7af 100644
--- a/llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll
+++ b/llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll
@@ -1,6 +1,6 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512bw --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512bw --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mcpu=generic -mattr=+avx512bw --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=generic -mattr=+avx512bw --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64
declare i32 @llvm.x86.avx512.kunpck.wd(i32, i32)
@@ -1908,16 +1908,17 @@ define i64 @test_cmp_b_512(<64 x i8> %a0, <64 x i8> %a1) nounwind {
; X64-NEXT: addq %rax, %rcx # encoding: [0x48,0x01,0xc1]
; X64-NEXT: vpcmpleb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3f,0xc1,0x02]
; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
+; X64-NEXT: addq %rcx, %rax # encoding: [0x48,0x01,0xc8]
; X64-NEXT: vpcmpneqb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3f,0xc1,0x04]
+; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
+; X64-NEXT: vpcmpnltb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3f,0xc1,0x05]
; X64-NEXT: kmovq %k0, %rdx # encoding: [0xc4,0xe1,0xfb,0x93,0xd0]
-; X64-NEXT: addq %rax, %rdx # encoding: [0x48,0x01,0xc2]
; X64-NEXT: addq %rcx, %rdx # encoding: [0x48,0x01,0xca]
-; X64-NEXT: vpcmpnltb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3f,0xc1,0x05]
+; X64-NEXT: addq %rax, %rdx # encoding: [0x48,0x01,0xc2]
+; X64-NEXT: vpcmpgtb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf1,0x7d,0x48,0x64,0xc1]
; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
; X64-NEXT: addq %rdx, %rax # encoding: [0x48,0x01,0xd0]
-; X64-NEXT: vpcmpgtb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf1,0x7d,0x48,0x64,0xc1]
-; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
-; X64-NEXT: leaq -1(%rcx,%rax), %rax # encoding: [0x48,0x8d,0x44,0x01,0xff]
+; X64-NEXT: decq %rax # encoding: [0x48,0xff,0xc8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i64 @llvm.x86.avx512.mask.cmp.b.512(<64 x i8> %a0, <64 x i8> %a1, i32 0, i64 -1)
@@ -1994,18 +1995,18 @@ define i64 @test_mask_cmp_b_512(<64 x i8> %a0, <64 x i8> %a1, i64 %mask) nounwin
; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
; X64-NEXT: addq %rax, %rcx # encoding: [0x48,0x01,0xc1]
; X64-NEXT: vpcmpleb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3f,0xc1,0x02]
-; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
-; X64-NEXT: vpcmpneqb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3f,0xc1,0x04]
; X64-NEXT: kmovq %k0, %rdx # encoding: [0xc4,0xe1,0xfb,0x93,0xd0]
-; X64-NEXT: addq %rax, %rdx # encoding: [0x48,0x01,0xc2]
; X64-NEXT: addq %rcx, %rdx # encoding: [0x48,0x01,0xca]
+; X64-NEXT: vpcmpneqb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3f,0xc1,0x04]
+; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
; X64-NEXT: vpcmpnltb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3f,0xc1,0x05]
; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
+; X64-NEXT: addq %rax, %rcx # encoding: [0x48,0x01,0xc1]
; X64-NEXT: vpcmpgtb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0x64,0xc1]
; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
; X64-NEXT: addq %rcx, %rax # encoding: [0x48,0x01,0xc8]
-; X64-NEXT: addq %rdi, %rax # encoding: [0x48,0x01,0xf8]
; X64-NEXT: addq %rdx, %rax # encoding: [0x48,0x01,0xd0]
+; X64-NEXT: addq %rdi, %rax # encoding: [0x48,0x01,0xf8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i64 @llvm.x86.avx512.mask.cmp.b.512(<64 x i8> %a0, <64 x i8> %a1, i32 0, i64 %mask)
@@ -2083,16 +2084,17 @@ define i64 @test_ucmp_b_512(<64 x i8> %a0, <64 x i8> %a1) nounwind {
; X64-NEXT: addq %rax, %rcx # encoding: [0x48,0x01,0xc1]
; X64-NEXT: vpcmpleub %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3e,0xc1,0x02]
; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
+; X64-NEXT: addq %rcx, %rax # encoding: [0x48,0x01,0xc8]
; X64-NEXT: vpcmpneqb %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3f,0xc1,0x04]
+; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
+; X64-NEXT: vpcmpnltub %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3e,0xc1,0x05]
; X64-NEXT: kmovq %k0, %rdx # encoding: [0xc4,0xe1,0xfb,0x93,0xd0]
-; X64-NEXT: addq %rax, %rdx # encoding: [0x48,0x01,0xc2]
; X64-NEXT: addq %rcx, %rdx # encoding: [0x48,0x01,0xca]
-; X64-NEXT: vpcmpnltub %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3e,0xc1,0x05]
+; X64-NEXT: addq %rax, %rdx # encoding: [0x48,0x01,0xc2]
+; X64-NEXT: vpcmpnleub %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3e,0xc1,0x06]
; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
; X64-NEXT: addq %rdx, %rax # encoding: [0x48,0x01,0xd0]
-; X64-NEXT: vpcmpnleub %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0x7d,0x48,0x3e,0xc1,0x06]
-; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
-; X64-NEXT: leaq -1(%rcx,%rax), %rax # encoding: [0x48,0x8d,0x44,0x01,0xff]
+; X64-NEXT: decq %rax # encoding: [0x48,0xff,0xc8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i64 @llvm.x86.avx512.mask.ucmp.b.512(<64 x i8> %a0, <64 x i8> %a1, i32 0, i64 -1)
@@ -2169,18 +2171,18 @@ define i64 @test_mask_x86_avx512_ucmp_b_512(<64 x i8> %a0, <64 x i8> %a1, i64 %m
; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
; X64-NEXT: addq %rax, %rcx # encoding: [0x48,0x01,0xc1]
; X64-NEXT: vpcmpleub %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3e,0xc1,0x02]
-; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
-; X64-NEXT: vpcmpneqb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3f,0xc1,0x04]
; X64-NEXT: kmovq %k0, %rdx # encoding: [0xc4,0xe1,0xfb,0x93,0xd0]
-; X64-NEXT: addq %rax, %rdx # encoding: [0x48,0x01,0xc2]
; X64-NEXT: addq %rcx, %rdx # encoding: [0x48,0x01,0xca]
+; X64-NEXT: vpcmpneqb %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3f,0xc1,0x04]
+; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
; X64-NEXT: vpcmpnltub %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3e,0xc1,0x05]
; X64-NEXT: kmovq %k0, %rcx # encoding: [0xc4,0xe1,0xfb,0x93,0xc8]
+; X64-NEXT: addq %rax, %rcx # encoding: [0x48,0x01,0xc1]
; X64-NEXT: vpcmpnleub %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x3e,0xc1,0x06]
; X64-NEXT: kmovq %k0, %rax # encoding: [0xc4,0xe1,0xfb,0x93,0xc0]
; X64-NEXT: addq %rcx, %rax # encoding: [0x48,0x01,0xc8]
-; X64-NEXT: addq %rdi, %rax # encoding: [0x48,0x01,0xf8]
; X64-NEXT: addq %rdx, %rax # encoding: [0x48,0x01,0xd0]
+; X64-NEXT: addq %rdi, %rax # encoding: [0x48,0x01,0xf8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i64 @llvm.x86.avx512.mask.ucmp.b.512(<64 x i8> %a0, <64 x i8> %a1, i32 0, i64 %mask)
@@ -2213,16 +2215,17 @@ define i32 @test_cmp_w_512(<32 x i16> %a0, <32 x i16> %a1) nounwind {
; X86-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X86-NEXT: vpcmplew %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x02]
; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
+; X86-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X86-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x04]
+; X86-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
+; X86-NEXT: vpcmpnltw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x05]
; X86-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
-; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X86-NEXT: addl %ecx, %edx # encoding: [0x01,0xca]
-; X86-NEXT: vpcmpnltw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x05]
+; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
+; X86-NEXT: vpcmpgtw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf1,0x7d,0x48,0x65,0xc1]
; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X86-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
-; X86-NEXT: vpcmpgtw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf1,0x7d,0x48,0x65,0xc1]
-; X86-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
-; X86-NEXT: leal -1(%ecx,%eax), %eax # encoding: [0x8d,0x44,0x01,0xff]
+; X86-NEXT: decl %eax # encoding: [0x48]
; X86-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -2235,16 +2238,17 @@ define i32 @test_cmp_w_512(<32 x i16> %a0, <32 x i16> %a1) nounwind {
; X64-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X64-NEXT: vpcmplew %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x02]
; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
+; X64-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X64-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x04]
+; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
+; X64-NEXT: vpcmpnltw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x05]
; X64-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
-; X64-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X64-NEXT: addl %ecx, %edx # encoding: [0x01,0xca]
-; X64-NEXT: vpcmpnltw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x05]
+; X64-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
+; X64-NEXT: vpcmpgtw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf1,0x7d,0x48,0x65,0xc1]
; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X64-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
-; X64-NEXT: vpcmpgtw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf1,0x7d,0x48,0x65,0xc1]
-; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
-; X64-NEXT: leal -1(%rcx,%rax), %eax # encoding: [0x8d,0x44,0x01,0xff]
+; X64-NEXT: decl %eax # encoding: [0xff,0xc8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.cmp.w.512(<32 x i16> %a0, <32 x i16> %a1, i32 0, i32 -1)
@@ -2277,18 +2281,18 @@ define i32 @test_mask_cmp_w_512(<32 x i16> %a0, <32 x i16> %a1, i32 %mask) nounw
; X86-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X86-NEXT: vpcmplew %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x02]
-; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
-; X86-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
; X86-NEXT: kmovd %k0, %esi # encoding: [0xc5,0xfb,0x93,0xf0]
-; X86-NEXT: addl %eax, %esi # encoding: [0x01,0xc6]
; X86-NEXT: addl %edx, %esi # encoding: [0x01,0xd6]
+; X86-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
+; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X86-NEXT: vpcmpnltw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x05]
; X86-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
+; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X86-NEXT: vpcmpgtw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0x65,0xc1]
; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X86-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
-; X86-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X86-NEXT: addl %esi, %eax # encoding: [0x01,0xf0]
+; X86-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X86-NEXT: popl %esi # encoding: [0x5e]
; X86-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT: retl # encoding: [0xc3]
@@ -2302,18 +2306,18 @@ define i32 @test_mask_cmp_w_512(<32 x i16> %a0, <32 x i16> %a1, i32 %mask) nounw
; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
; X64-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X64-NEXT: vpcmplew %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x02]
-; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
-; X64-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
; X64-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
-; X64-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X64-NEXT: addl %ecx, %edx # encoding: [0x01,0xca]
+; X64-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
+; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X64-NEXT: vpcmpnltw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x05]
; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
+; X64-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X64-NEXT: vpcmpgtw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0x65,0xc1]
; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X64-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
-; X64-NEXT: addl %edi, %eax # encoding: [0x01,0xf8]
; X64-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
+; X64-NEXT: addl %edi, %eax # encoding: [0x01,0xf8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.cmp.w.512(<32 x i16> %a0, <32 x i16> %a1, i32 0, i32 %mask)
@@ -2346,16 +2350,17 @@ define i32 @test_ucmp_w_512(<32 x i16> %a0, <32 x i16> %a1) nounwind {
; X86-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X86-NEXT: vpcmpleuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x02]
; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
+; X86-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X86-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x04]
+; X86-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
+; X86-NEXT: vpcmpnltuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x05]
; X86-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
-; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X86-NEXT: addl %ecx, %edx # encoding: [0x01,0xca]
-; X86-NEXT: vpcmpnltuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x05]
+; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
+; X86-NEXT: vpcmpnleuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x06]
; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X86-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
-; X86-NEXT: vpcmpnleuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x06]
-; X86-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
-; X86-NEXT: leal -1(%ecx,%eax), %eax # encoding: [0x8d,0x44,0x01,0xff]
+; X86-NEXT: decl %eax # encoding: [0x48]
; X86-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT: retl # encoding: [0xc3]
;
@@ -2368,16 +2373,17 @@ define i32 @test_ucmp_w_512(<32 x i16> %a0, <32 x i16> %a1) nounwind {
; X64-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X64-NEXT: vpcmpleuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x02]
; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
+; X64-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X64-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3f,0xc1,0x04]
+; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
+; X64-NEXT: vpcmpnltuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x05]
; X64-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
-; X64-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X64-NEXT: addl %ecx, %edx # encoding: [0x01,0xca]
-; X64-NEXT: vpcmpnltuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x05]
+; X64-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
+; X64-NEXT: vpcmpnleuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x06]
; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X64-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
-; X64-NEXT: vpcmpnleuw %zmm1, %zmm0, %k0 # encoding: [0x62,0xf3,0xfd,0x48,0x3e,0xc1,0x06]
-; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
-; X64-NEXT: leal -1(%rcx,%rax), %eax # encoding: [0x8d,0x44,0x01,0xff]
+; X64-NEXT: decl %eax # encoding: [0xff,0xc8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.ucmp.w.512(<32 x i16> %a0, <32 x i16> %a1, i32 0, i32 -1)
@@ -2410,18 +2416,18 @@ define i32 @test_mask_ucmp_w_512(<32 x i16> %a0, <32 x i16> %a1, i32 %mask) noun
; X86-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X86-NEXT: vpcmpleuw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3e,0xc1,0x02]
-; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
-; X86-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
; X86-NEXT: kmovd %k0, %esi # encoding: [0xc5,0xfb,0x93,0xf0]
-; X86-NEXT: addl %eax, %esi # encoding: [0x01,0xc6]
; X86-NEXT: addl %edx, %esi # encoding: [0x01,0xd6]
+; X86-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
+; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X86-NEXT: vpcmpnltuw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3e,0xc1,0x05]
; X86-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
+; X86-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X86-NEXT: vpcmpnleuw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3e,0xc1,0x06]
; X86-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X86-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
-; X86-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X86-NEXT: addl %esi, %eax # encoding: [0x01,0xf0]
+; X86-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
; X86-NEXT: popl %esi # encoding: [0x5e]
; X86-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X86-NEXT: retl # encoding: [0xc3]
@@ -2435,18 +2441,18 @@ define i32 @test_mask_ucmp_w_512(<32 x i16> %a0, <32 x i16> %a1, i32 %mask) noun
; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
; X64-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X64-NEXT: vpcmpleuw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3e,0xc1,0x02]
-; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
-; X64-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
; X64-NEXT: kmovd %k0, %edx # encoding: [0xc5,0xfb,0x93,0xd0]
-; X64-NEXT: addl %eax, %edx # encoding: [0x01,0xc2]
; X64-NEXT: addl %ecx, %edx # encoding: [0x01,0xca]
+; X64-NEXT: vpcmpneqw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3f,0xc1,0x04]
+; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X64-NEXT: vpcmpnltuw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3e,0xc1,0x05]
; X64-NEXT: kmovd %k0, %ecx # encoding: [0xc5,0xfb,0x93,0xc8]
+; X64-NEXT: addl %eax, %ecx # encoding: [0x01,0xc1]
; X64-NEXT: vpcmpnleuw %zmm1, %zmm0, %k0 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x3e,0xc1,0x06]
; X64-NEXT: kmovd %k0, %eax # encoding: [0xc5,0xfb,0x93,0xc0]
; X64-NEXT: addl %ecx, %eax # encoding: [0x01,0xc8]
-; X64-NEXT: addl %edi, %eax # encoding: [0x01,0xf8]
; X64-NEXT: addl %edx, %eax # encoding: [0x01,0xd0]
+; X64-NEXT: addl %edi, %eax # encoding: [0x01,0xf8]
; X64-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-NEXT: retq # encoding: [0xc3]
%res0 = call i32 @llvm.x86.av...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/111865
More information about the llvm-commits
mailing list