CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
182 lines
6.9 KiB
LLVM
182 lines
6.9 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=i686-- -mattr=+sse2,+pclmul | FileCheck %s
|
|
|
|
; Minimal test for i686 PCLMULQDQ lowering
|
|
|
|
define i32 @clmul_i32(i32 %a, i32 %b) nounwind {
|
|
; CHECK-LABEL: clmul_i32:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
|
|
; CHECK-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero
|
|
; CHECK-NEXT: pclmulqdq $0, %xmm0, %xmm1
|
|
; CHECK-NEXT: movd %xmm1, %eax
|
|
; CHECK-NEXT: retl
|
|
%res = call i32 @llvm.clmul.i32(i32 %a, i32 %b)
|
|
ret i32 %res
|
|
}
|
|
|
|
define i64 @clmul_i64(i64 %a, i64 %b) nounwind {
|
|
; CHECK-LABEL: clmul_i64:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movq {{.*#+}} xmm0 = mem[0],zero
|
|
; CHECK-NEXT: movq {{.*#+}} xmm1 = mem[0],zero
|
|
; CHECK-NEXT: pclmulqdq $0, %xmm0, %xmm1
|
|
; CHECK-NEXT: movd %xmm1, %eax
|
|
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
|
|
; CHECK-NEXT: movd %xmm0, %edx
|
|
; CHECK-NEXT: retl
|
|
%res = call i64 @llvm.clmul.i64(i64 %a, i64 %b)
|
|
ret i64 %res
|
|
}
|
|
|
|
define i64 @clmulr_i64(i64 %a, i64 %b) nounwind {
|
|
; CHECK-LABEL: clmulr_i64:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: pushl %edi
|
|
; CHECK-NEXT: pushl %esi
|
|
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
|
|
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
|
|
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %esi
|
|
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %edx
|
|
; CHECK-NEXT: bswapl %esi
|
|
; CHECK-NEXT: movl %esi, %edi
|
|
; CHECK-NEXT: andl $252645135, %edi # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: shll $4, %edi
|
|
; CHECK-NEXT: shrl $4, %esi
|
|
; CHECK-NEXT: andl $252645135, %esi # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: orl %edi, %esi
|
|
; CHECK-NEXT: movl %esi, %edi
|
|
; CHECK-NEXT: andl $858993459, %edi # imm = 0x33333333
|
|
; CHECK-NEXT: shrl $2, %esi
|
|
; CHECK-NEXT: andl $858993459, %esi # imm = 0x33333333
|
|
; CHECK-NEXT: leal (%esi,%edi,4), %esi
|
|
; CHECK-NEXT: movl %esi, %edi
|
|
; CHECK-NEXT: andl $1431655765, %edi # imm = 0x55555555
|
|
; CHECK-NEXT: shrl %esi
|
|
; CHECK-NEXT: andl $1431655765, %esi # imm = 0x55555555
|
|
; CHECK-NEXT: leal (%esi,%edi,2), %esi
|
|
; CHECK-NEXT: movd %esi, %xmm1
|
|
; CHECK-NEXT: bswapl %edx
|
|
; CHECK-NEXT: movl %edx, %esi
|
|
; CHECK-NEXT: andl $252645135, %esi # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: shll $4, %esi
|
|
; CHECK-NEXT: shrl $4, %edx
|
|
; CHECK-NEXT: andl $252645135, %edx # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: orl %esi, %edx
|
|
; CHECK-NEXT: movl %edx, %esi
|
|
; CHECK-NEXT: andl $858993459, %esi # imm = 0x33333333
|
|
; CHECK-NEXT: shrl $2, %edx
|
|
; CHECK-NEXT: andl $858993459, %edx # imm = 0x33333333
|
|
; CHECK-NEXT: leal (%edx,%esi,4), %edx
|
|
; CHECK-NEXT: movl %edx, %esi
|
|
; CHECK-NEXT: andl $1431655765, %esi # imm = 0x55555555
|
|
; CHECK-NEXT: shrl %edx
|
|
; CHECK-NEXT: andl $1431655765, %edx # imm = 0x55555555
|
|
; CHECK-NEXT: leal (%edx,%esi,2), %edx
|
|
; CHECK-NEXT: movd %edx, %xmm0
|
|
; CHECK-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
|
|
; CHECK-NEXT: bswapl %ecx
|
|
; CHECK-NEXT: movl %ecx, %edx
|
|
; CHECK-NEXT: andl $252645135, %edx # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: shll $4, %edx
|
|
; CHECK-NEXT: shrl $4, %ecx
|
|
; CHECK-NEXT: andl $252645135, %ecx # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: orl %edx, %ecx
|
|
; CHECK-NEXT: movl %ecx, %edx
|
|
; CHECK-NEXT: andl $858993459, %edx # imm = 0x33333333
|
|
; CHECK-NEXT: shrl $2, %ecx
|
|
; CHECK-NEXT: andl $858993459, %ecx # imm = 0x33333333
|
|
; CHECK-NEXT: leal (%ecx,%edx,4), %ecx
|
|
; CHECK-NEXT: movl %ecx, %edx
|
|
; CHECK-NEXT: andl $1431655765, %edx # imm = 0x55555555
|
|
; CHECK-NEXT: shrl %ecx
|
|
; CHECK-NEXT: andl $1431655765, %ecx # imm = 0x55555555
|
|
; CHECK-NEXT: leal (%ecx,%edx,2), %ecx
|
|
; CHECK-NEXT: movd %ecx, %xmm2
|
|
; CHECK-NEXT: bswapl %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $252645135, %ecx # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: shll $4, %ecx
|
|
; CHECK-NEXT: shrl $4, %eax
|
|
; CHECK-NEXT: andl $252645135, %eax # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: orl %ecx, %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $858993459, %ecx # imm = 0x33333333
|
|
; CHECK-NEXT: shrl $2, %eax
|
|
; CHECK-NEXT: andl $858993459, %eax # imm = 0x33333333
|
|
; CHECK-NEXT: leal (%eax,%ecx,4), %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $1431655765, %ecx # imm = 0x55555555
|
|
; CHECK-NEXT: shrl %eax
|
|
; CHECK-NEXT: andl $1431655765, %eax # imm = 0x55555555
|
|
; CHECK-NEXT: leal (%eax,%ecx,2), %eax
|
|
; CHECK-NEXT: movd %eax, %xmm1
|
|
; CHECK-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
|
|
; CHECK-NEXT: pclmulqdq $0, %xmm0, %xmm1
|
|
; CHECK-NEXT: movd %xmm1, %eax
|
|
; CHECK-NEXT: bswapl %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $252645135, %ecx # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: shll $4, %ecx
|
|
; CHECK-NEXT: shrl $4, %eax
|
|
; CHECK-NEXT: andl $252645135, %eax # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: orl %ecx, %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $858993459, %ecx # imm = 0x33333333
|
|
; CHECK-NEXT: shrl $2, %eax
|
|
; CHECK-NEXT: andl $858993459, %eax # imm = 0x33333333
|
|
; CHECK-NEXT: leal (%eax,%ecx,4), %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $1431655765, %ecx # imm = 0x55555555
|
|
; CHECK-NEXT: shrl %eax
|
|
; CHECK-NEXT: andl $1431655765, %eax # imm = 0x55555555
|
|
; CHECK-NEXT: leal (%eax,%ecx,2), %edx
|
|
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
|
|
; CHECK-NEXT: movd %xmm0, %eax
|
|
; CHECK-NEXT: bswapl %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $252645135, %ecx # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: shll $4, %ecx
|
|
; CHECK-NEXT: shrl $4, %eax
|
|
; CHECK-NEXT: andl $252645135, %eax # imm = 0xF0F0F0F
|
|
; CHECK-NEXT: orl %ecx, %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $858993459, %ecx # imm = 0x33333333
|
|
; CHECK-NEXT: shrl $2, %eax
|
|
; CHECK-NEXT: andl $858993459, %eax # imm = 0x33333333
|
|
; CHECK-NEXT: leal (%eax,%ecx,4), %eax
|
|
; CHECK-NEXT: movl %eax, %ecx
|
|
; CHECK-NEXT: andl $1431655765, %ecx # imm = 0x55555555
|
|
; CHECK-NEXT: shrl %eax
|
|
; CHECK-NEXT: andl $1431655765, %eax # imm = 0x55555555
|
|
; CHECK-NEXT: leal (%eax,%ecx,2), %eax
|
|
; CHECK-NEXT: popl %esi
|
|
; CHECK-NEXT: popl %edi
|
|
; CHECK-NEXT: retl
|
|
%a.ext = zext i64 %a to i128
|
|
%b.ext = zext i64 %b to i128
|
|
%clmul = call i128 @llvm.clmul.i128(i128 %a.ext, i128 %b.ext)
|
|
%res.ext = lshr i128 %clmul, 63
|
|
%res = trunc i128 %res.ext to i64
|
|
ret i64 %res
|
|
}
|
|
|
|
define i64 @clmulh_i64(i64 %a, i64 %b) nounwind {
|
|
; CHECK-LABEL: clmulh_i64:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movq {{.*#+}} xmm0 = mem[0],zero
|
|
; CHECK-NEXT: movq {{.*#+}} xmm1 = mem[0],zero
|
|
; CHECK-NEXT: pclmulqdq $0, %xmm0, %xmm1
|
|
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
|
|
; CHECK-NEXT: movd %xmm0, %eax
|
|
; CHECK-NEXT: pshufd {{.*#+}} xmm0 = xmm1[3,3,3,3]
|
|
; CHECK-NEXT: movd %xmm0, %edx
|
|
; CHECK-NEXT: retl
|
|
%a.ext = zext i64 %a to i128
|
|
%b.ext = zext i64 %b to i128
|
|
%clmul = call i128 @llvm.clmul.i128(i128 %a.ext, i128 %b.ext)
|
|
%res.ext = lshr i128 %clmul, 64
|
|
%res = trunc i128 %res.ext to i64
|
|
ret i64 %res
|
|
}
|