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
75 lines
2.9 KiB
LLVM
75 lines
2.9 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
|
|
; RUN: llc -mtriple=aarch64-linux -mattr=+neon,+fp8dot2,+fp8dot4 < %s | FileCheck %s
|
|
|
|
define <4 x half> @test_fdot_f16(<4 x half> %vd, <8 x i8> %vn, <8 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdot_f16:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.4h, v1.8b, v2.8b
|
|
; CHECK-NEXT: ret
|
|
%res = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.v4f16.v8i8(<4 x half> %vd, <8 x i8> %vn, <8 x i8> %vm)
|
|
ret <4 x half> %res
|
|
}
|
|
|
|
define <8 x half> @test_fdotq_f16(<8 x half> %vd, <16 x i8> %vn, <16 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdotq_f16:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.8h, v1.16b, v2.16b
|
|
; CHECK-NEXT: ret
|
|
%res = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.v8f16.v16i8(<8 x half> %vd, <16 x i8> %vn, <16 x i8> %vm)
|
|
ret <8 x half> %res
|
|
}
|
|
|
|
define <4 x half> @test_fdot_lane_f16(<4 x half> %vd, <8 x i8> %vn, <16 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdot_lane_f16:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.4h, v1.8b, v2.2b[0]
|
|
; CHECK-NEXT: ret
|
|
%res = call <4 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v4f16.v8i8(<4 x half> %vd, <8 x i8> %vn, <16 x i8> %vm, i32 0)
|
|
ret <4 x half> %res
|
|
}
|
|
|
|
define <8 x half> @test_fdotq_lane_f16(<8 x half> %vd, <16 x i8> %vn, <16 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdotq_lane_f16:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.8h, v1.16b, v2.2b[7]
|
|
; CHECK-NEXT: ret
|
|
%res = call <8 x half> @llvm.aarch64.neon.fp8.fdot2.lane.v8f16.v16i8(<8 x half> %vd, <16 x i8> %vn, <16 x i8> %vm, i32 7)
|
|
ret <8 x half> %res
|
|
}
|
|
|
|
define <2 x float> @test_fdot_f32(<2 x float> %vd, <8 x i8> %vn, <8 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdot_f32:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.2s, v1.8b, v2.8b
|
|
; CHECK-NEXT: ret
|
|
%res = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.v2f32.v8i8(<2 x float> %vd, <8 x i8> %vn, <8 x i8> %vm)
|
|
ret <2 x float> %res
|
|
}
|
|
|
|
define <4 x float> @test_fdotq_f32(<4 x float> %vd, <16 x i8> %vn, <16 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdotq_f32:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.4s, v1.16b, v2.16b
|
|
; CHECK-NEXT: ret
|
|
%res = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.v4f32.v16i8(<4 x float> %vd, <16 x i8> %vn, <16 x i8> %vm)
|
|
ret <4 x float> %res
|
|
}
|
|
|
|
define <2 x float> @test_fdot_lane_f32(<2 x float> %vd, <8 x i8> %vn, <16 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdot_lane_f32:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.2s, v1.8b, v2.4b[0]
|
|
; CHECK-NEXT: ret
|
|
%res = call <2 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v2f32.v8i8(<2 x float> %vd, <8 x i8> %vn, <16 x i8> %vm, i32 0)
|
|
ret <2 x float> %res
|
|
}
|
|
|
|
define <4 x float> @test_fdotq_lane_f32(<4 x float> %vd, <16 x i8> %vn, <16 x i8> %vm) {
|
|
; CHECK-LABEL: test_fdotq_lane_f32:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: fdot v0.4s, v1.16b, v2.4b[3]
|
|
; CHECK-NEXT: ret
|
|
%res = call <4 x float> @llvm.aarch64.neon.fp8.fdot4.lane.v4f32.v16i8(<4 x float> %vd, <16 x i8> %vn, <16 x i8> %vm, i32 3)
|
|
ret <4 x float> %res
|
|
}
|