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
91 lines
3.7 KiB
LLVM
91 lines
3.7 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc -force-streaming -verify-machineinstrs < %s | FileCheck %s
|
|
|
|
target triple = "aarch64-linux"
|
|
|
|
define void @mop4a_za16_fp8_1x1(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm) #0 {
|
|
; CHECK-LABEL: mop4a_za16_fp8_1x1:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z24.d, z1.d
|
|
; CHECK-NEXT: fmop4a za0.h, z0.b, z24.b
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za16.1x1(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za16_fp8_1x2(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
|
|
; CHECK-LABEL: mop4a_za16_fp8_1x2:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z25.d, z2.d
|
|
; CHECK-NEXT: mov z24.d, z1.d
|
|
; CHECK-NEXT: fmop4a za0.h, z0.b, { z24.b, z25.b }
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za16.1x2(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za16_fp8_2x1(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm) #0 {
|
|
; CHECK-LABEL: mop4a_za16_fp8_2x1:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z24.d, z2.d
|
|
; CHECK-NEXT: fmop4a za0.h, { z0.b, z1.b }, z24.b
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za16.2x1(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za16_fp8_2x2(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
|
|
; CHECK-LABEL: mop4a_za16_fp8_2x2:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z25.d, z3.d
|
|
; CHECK-NEXT: mov z24.d, z2.d
|
|
; CHECK-NEXT: fmop4a za0.h, { z0.b, z1.b }, { z24.b, z25.b }
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za16.2x2(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za32_fp8(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm) #0 {
|
|
; CHECK-LABEL: mop4a_za32_fp8:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z24.d, z1.d
|
|
; CHECK-NEXT: fmop4a za0.s, z0.b, z24.b
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za32.1x1(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za32_fp8_1x2(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
|
|
; CHECK-LABEL: mop4a_za32_fp8_1x2:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z25.d, z2.d
|
|
; CHECK-NEXT: mov z24.d, z1.d
|
|
; CHECK-NEXT: fmop4a za0.s, z0.b, { z24.b, z25.b }
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za32.1x2(i32 0, <vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za32_fp8_2x1(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm) #0 {
|
|
; CHECK-LABEL: mop4a_za32_fp8_2x1:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z24.d, z2.d
|
|
; CHECK-NEXT: fmop4a za0.s, { z0.b, z1.b }, z24.b
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za32.2x1(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm)
|
|
ret void
|
|
}
|
|
|
|
define void @mop4a_za32_fp8_2x2(<vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2) #0 {
|
|
; CHECK-LABEL: mop4a_za32_fp8_2x2:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: mov z25.d, z3.d
|
|
; CHECK-NEXT: mov z24.d, z2.d
|
|
; CHECK-NEXT: fmop4a za0.s, { z0.b, z1.b }, { z24.b, z25.b }
|
|
; CHECK-NEXT: ret
|
|
call void @llvm.aarch64.sme.fp8.fmop4a.za32.2x2(i32 0, <vscale x 16 x i8> %zn1, <vscale x 16 x i8> %zn2, <vscale x 16 x i8> %zm1, <vscale x 16 x i8> %zm2)
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = {nounwind "target-features" = "+sme-f8f16,+sme-f8f32,+sme2p1,+sme-mop4" }
|