derek.holloway
7502bf64a4
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
134 lines
6.7 KiB
LLVM
134 lines
6.7 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avxvnni --show-mc-encoding | FileCheck %s --check-prefixes=AVXVNNI
|
|
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512vnni,+avx512vl,+avxvnni --show-mc-encoding | FileCheck %s --check-prefixes=AVX512VNNI
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avxvnni --show-mc-encoding | FileCheck %s --check-prefixes=AVXVNNI
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vnni,+avx512vl,+avxvnni --show-mc-encoding | FileCheck %s --check-prefixes=AVX512VNNI
|
|
|
|
declare <8 x i32> @llvm.x86.avx512.vpdpbusd.256(<8 x i32>, <32 x i8>, <32 x i8>)
|
|
|
|
define <8 x i32>@test_int_x86_avx_vpdpbusd_256(<8 x i32> %x0, <32 x i8> %x1, <32 x i8> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpbusd_256:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpbusd %ymm2, %ymm1, %ymm0 # encoding: [0xc4,0xe2,0x75,0x50,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpbusd_256:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpbusd %ymm2, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x75,0x50,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <8 x i32> @llvm.x86.avx512.vpdpbusd.256(<8 x i32> %x0, <32 x i8> %x1, <32 x i8> %x2)
|
|
ret <8 x i32> %res
|
|
}
|
|
|
|
declare <4 x i32> @llvm.x86.avx512.vpdpbusd.128(<4 x i32>, <16 x i8>, <16 x i8>)
|
|
|
|
define <4 x i32>@test_int_x86_avx_vpdpbusd_128(<4 x i32> %x0, <16 x i8> %x1, <16 x i8> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpbusd_128:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpbusd %xmm2, %xmm1, %xmm0 # encoding: [0xc4,0xe2,0x71,0x50,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpbusd_128:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpbusd %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x71,0x50,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <4 x i32> @llvm.x86.avx512.vpdpbusd.128(<4 x i32> %x0, <16 x i8> %x1, <16 x i8> %x2)
|
|
ret <4 x i32> %res
|
|
}
|
|
|
|
declare <8 x i32> @llvm.x86.avx512.vpdpbusds.256(<8 x i32>, <32 x i8>, <32 x i8>)
|
|
|
|
define <8 x i32>@test_int_x86_avx_vpdpbusds_256(<8 x i32> %x0, <32 x i8> %x1, <32 x i8> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpbusds_256:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpbusds %ymm2, %ymm1, %ymm0 # encoding: [0xc4,0xe2,0x75,0x51,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpbusds_256:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpbusds %ymm2, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x75,0x51,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <8 x i32> @llvm.x86.avx512.vpdpbusds.256(<8 x i32> %x0, <32 x i8> %x1, <32 x i8> %x2)
|
|
ret <8 x i32> %res
|
|
}
|
|
|
|
declare <4 x i32> @llvm.x86.avx512.vpdpbusds.128(<4 x i32>, <16 x i8>, <16 x i8>)
|
|
|
|
define <4 x i32>@test_int_x86_avx_vpdpbusds_128(<4 x i32> %x0, <16 x i8> %x1, <16 x i8> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpbusds_128:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpbusds %xmm2, %xmm1, %xmm0 # encoding: [0xc4,0xe2,0x71,0x51,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpbusds_128:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpbusds %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x71,0x51,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <4 x i32> @llvm.x86.avx512.vpdpbusds.128(<4 x i32> %x0, <16 x i8> %x1, <16 x i8> %x2)
|
|
ret <4 x i32> %res
|
|
}
|
|
|
|
declare <8 x i32> @llvm.x86.avx512.vpdpwssd.256(<8 x i32>, <16 x i16>, <16 x i16>)
|
|
|
|
define <8 x i32>@test_int_x86_avx_vpdpwssd_256(<8 x i32> %x0, <16 x i16> %x1, <16 x i16> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpwssd_256:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpwssd %ymm2, %ymm1, %ymm0 # encoding: [0xc4,0xe2,0x75,0x52,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpwssd_256:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpwssd %ymm2, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x75,0x52,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <8 x i32> @llvm.x86.avx512.vpdpwssd.256(<8 x i32> %x0, <16 x i16> %x1, <16 x i16> %x2)
|
|
ret <8 x i32> %res
|
|
}
|
|
|
|
declare <4 x i32> @llvm.x86.avx512.vpdpwssd.128(<4 x i32>, <8 x i16>, <8 x i16>)
|
|
|
|
define <4 x i32>@test_int_x86_avx_vpdpwssd_128(<4 x i32> %x0, <8 x i16> %x1, <8 x i16> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpwssd_128:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpwssd %xmm2, %xmm1, %xmm0 # encoding: [0xc4,0xe2,0x71,0x52,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpwssd_128:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpwssd %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x71,0x52,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <4 x i32> @llvm.x86.avx512.vpdpwssd.128(<4 x i32> %x0, <8 x i16> %x1, <8 x i16> %x2)
|
|
ret <4 x i32> %res
|
|
}
|
|
|
|
declare <8 x i32> @llvm.x86.avx512.vpdpwssds.256(<8 x i32>, <16 x i16>, <16 x i16>)
|
|
|
|
define <8 x i32>@test_int_x86_avx_vpdpwssds_256(<8 x i32> %x0, <16 x i16> %x1, <16 x i16> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpwssds_256:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpwssds %ymm2, %ymm1, %ymm0 # encoding: [0xc4,0xe2,0x75,0x53,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpwssds_256:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpwssds %ymm2, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x75,0x53,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <8 x i32> @llvm.x86.avx512.vpdpwssds.256(<8 x i32> %x0, <16 x i16> %x1, <16 x i16> %x2)
|
|
ret <8 x i32> %res
|
|
}
|
|
|
|
declare <4 x i32> @llvm.x86.avx512.vpdpwssds.128(<4 x i32>, <8 x i16>, <8 x i16>)
|
|
|
|
define <4 x i32>@test_int_x86_avx_vpdpwssds_128(<4 x i32> %x0, <8 x i16> %x1, <8 x i16> %x2) {
|
|
; AVXVNNI-LABEL: test_int_x86_avx_vpdpwssds_128:
|
|
; AVXVNNI: # %bb.0:
|
|
; AVXVNNI-NEXT: {vex} vpdpwssds %xmm2, %xmm1, %xmm0 # encoding: [0xc4,0xe2,0x71,0x53,0xc2]
|
|
; AVXVNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
;
|
|
; AVX512VNNI-LABEL: test_int_x86_avx_vpdpwssds_128:
|
|
; AVX512VNNI: # %bb.0:
|
|
; AVX512VNNI-NEXT: {vex} vpdpwssds %xmm2, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x71,0x53,0xc2]
|
|
; AVX512VNNI-NEXT: ret{{[l|q]}} # encoding: [0xc3]
|
|
%res = call <4 x i32> @llvm.x86.avx512.vpdpwssds.128(<4 x i32> %x0, <8 x i16> %x1, <8 x i16> %x2)
|
|
ret <4 x i32> %res
|
|
}
|