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
216 lines
16 KiB
LLVM
216 lines
16 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 4
|
|
; RUN: llc < %s -global-isel -stop-after=irtranslator -mtriple=amdgcn -mcpu=gfx900 | FileCheck %s -check-prefixes=GFX9
|
|
|
|
; tests bf16 argument & return values lowering.
|
|
|
|
define <3 x bfloat> @v3bf16(<3 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v3bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(s16), [[UV1:%[0-9]+]]:_(s16), [[UV2:%[0-9]+]]:_(s16), [[UV3:%[0-9]+]]:_(s16) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<4 x s16>)
|
|
; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<3 x s16>) = G_BUILD_VECTOR [[UV]](s16), [[UV1]](s16), [[UV2]](s16)
|
|
; GFX9-NEXT: [[C:%[0-9]+]]:_(s16) = G_FCONSTANT bfloat 0.000000e+00
|
|
; GFX9-NEXT: [[BUILD_VECTOR1:%[0-9]+]]:_(<3 x s16>) = G_BUILD_VECTOR [[C]](s16), [[C]](s16), [[C]](s16)
|
|
; GFX9-NEXT: [[SHUF:%[0-9]+]]:_(<3 x s16>) = G_SHUFFLE_VECTOR [[BUILD_VECTOR]](<3 x s16>), [[BUILD_VECTOR1]], shufflemask(3, 1, 2)
|
|
; GFX9-NEXT: [[UV4:%[0-9]+]]:_(s16), [[UV5:%[0-9]+]]:_(s16), [[UV6:%[0-9]+]]:_(s16) = G_UNMERGE_VALUES [[SHUF]](<3 x s16>)
|
|
; GFX9-NEXT: [[DEF:%[0-9]+]]:_(s16) = G_IMPLICIT_DEF
|
|
; GFX9-NEXT: [[BUILD_VECTOR2:%[0-9]+]]:_(<4 x s16>) = G_BUILD_VECTOR [[UV4]](s16), [[UV5]](s16), [[UV6]](s16), [[DEF]](s16)
|
|
; GFX9-NEXT: [[UV7:%[0-9]+]]:_(<2 x s16>), [[UV8:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<4 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV7]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV8]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1
|
|
%res = shufflevector <3 x bfloat> %arg0, <3 x bfloat> zeroinitializer, <3 x i32> <i32 3, i32 1, i32 2>
|
|
ret <3 x bfloat> %res
|
|
}
|
|
|
|
define <4 x bfloat> @v4bf16(<4 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v4bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<4 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>)
|
|
; GFX9-NEXT: [[C:%[0-9]+]]:_(s16) = G_FCONSTANT bfloat 0.000000e+00
|
|
; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<4 x s16>) = G_BUILD_VECTOR [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16)
|
|
; GFX9-NEXT: [[SHUF:%[0-9]+]]:_(<4 x s16>) = G_SHUFFLE_VECTOR [[CONCAT_VECTORS]](<4 x s16>), [[BUILD_VECTOR]], shufflemask(3, 1, 2, 0)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[SHUF]](<4 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV1]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1
|
|
%res = shufflevector <4 x bfloat> %arg0, <4 x bfloat> zeroinitializer, <4 x i32> <i32 3, i32 1, i32 2, i32 0>
|
|
ret <4 x bfloat> %res
|
|
}
|
|
|
|
define <5 x bfloat> @v5bf16(<5 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v5bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[COPY2]](<2 x s16>)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(s16), [[UV1:%[0-9]+]]:_(s16), [[UV2:%[0-9]+]]:_(s16), [[UV3:%[0-9]+]]:_(s16), [[UV4:%[0-9]+]]:_(s16), [[UV5:%[0-9]+]]:_(s16) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<6 x s16>)
|
|
; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<5 x s16>) = G_BUILD_VECTOR [[UV]](s16), [[UV1]](s16), [[UV2]](s16), [[UV3]](s16), [[UV4]](s16)
|
|
; GFX9-NEXT: [[C:%[0-9]+]]:_(s16) = G_FCONSTANT bfloat 0.000000e+00
|
|
; GFX9-NEXT: [[BUILD_VECTOR1:%[0-9]+]]:_(<5 x s16>) = G_BUILD_VECTOR [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16)
|
|
; GFX9-NEXT: [[SHUF:%[0-9]+]]:_(<5 x s16>) = G_SHUFFLE_VECTOR [[BUILD_VECTOR]](<5 x s16>), [[BUILD_VECTOR1]], shufflemask(3, 1, 2, 0, 4)
|
|
; GFX9-NEXT: [[UV6:%[0-9]+]]:_(s16), [[UV7:%[0-9]+]]:_(s16), [[UV8:%[0-9]+]]:_(s16), [[UV9:%[0-9]+]]:_(s16), [[UV10:%[0-9]+]]:_(s16) = G_UNMERGE_VALUES [[SHUF]](<5 x s16>)
|
|
; GFX9-NEXT: [[DEF:%[0-9]+]]:_(s16) = G_IMPLICIT_DEF
|
|
; GFX9-NEXT: [[BUILD_VECTOR2:%[0-9]+]]:_(<6 x s16>) = G_BUILD_VECTOR [[UV6]](s16), [[UV7]](s16), [[UV8]](s16), [[UV9]](s16), [[UV10]](s16), [[DEF]](s16)
|
|
; GFX9-NEXT: [[UV11:%[0-9]+]]:_(<2 x s16>), [[UV12:%[0-9]+]]:_(<2 x s16>), [[UV13:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<6 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV11]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV12]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr2 = COPY [[UV13]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
|
|
%res = shufflevector <5 x bfloat> %arg0, <5 x bfloat> zeroinitializer, <5 x i32> <i32 3, i32 1, i32 2, i32 0, i32 4>
|
|
ret <5 x bfloat> %res
|
|
}
|
|
|
|
define <6 x bfloat> @v6bf16(<6 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v6bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<6 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[COPY2]](<2 x s16>)
|
|
; GFX9-NEXT: [[C:%[0-9]+]]:_(s16) = G_FCONSTANT bfloat 0.000000e+00
|
|
; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<6 x s16>) = G_BUILD_VECTOR [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16)
|
|
; GFX9-NEXT: [[SHUF:%[0-9]+]]:_(<6 x s16>) = G_SHUFFLE_VECTOR [[CONCAT_VECTORS]](<6 x s16>), [[BUILD_VECTOR]], shufflemask(3, 1, 2, 0, 4, 5)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[SHUF]](<6 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV1]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr2 = COPY [[UV2]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2
|
|
%res = shufflevector <6 x bfloat> %arg0, <6 x bfloat> zeroinitializer, <6 x i32> <i32 3, i32 1, i32 2, i32 0, i32 4, i32 5>
|
|
ret <6 x bfloat> %res
|
|
}
|
|
|
|
define <7 x bfloat> @v7bf16(<7 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v7bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
|
|
; GFX9-NEXT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<8 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(s16), [[UV1:%[0-9]+]]:_(s16), [[UV2:%[0-9]+]]:_(s16), [[UV3:%[0-9]+]]:_(s16), [[UV4:%[0-9]+]]:_(s16), [[UV5:%[0-9]+]]:_(s16), [[UV6:%[0-9]+]]:_(s16), [[UV7:%[0-9]+]]:_(s16) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<8 x s16>)
|
|
; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<7 x s16>) = G_BUILD_VECTOR [[UV]](s16), [[UV1]](s16), [[UV2]](s16), [[UV3]](s16), [[UV4]](s16), [[UV5]](s16), [[UV6]](s16)
|
|
; GFX9-NEXT: [[C:%[0-9]+]]:_(s16) = G_FCONSTANT bfloat 0.000000e+00
|
|
; GFX9-NEXT: [[BUILD_VECTOR1:%[0-9]+]]:_(<7 x s16>) = G_BUILD_VECTOR [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16)
|
|
; GFX9-NEXT: [[SHUF:%[0-9]+]]:_(<7 x s16>) = G_SHUFFLE_VECTOR [[BUILD_VECTOR]](<7 x s16>), [[BUILD_VECTOR1]], shufflemask(3, 1, 2, 0, 4, 5, 6)
|
|
; GFX9-NEXT: [[UV8:%[0-9]+]]:_(s16), [[UV9:%[0-9]+]]:_(s16), [[UV10:%[0-9]+]]:_(s16), [[UV11:%[0-9]+]]:_(s16), [[UV12:%[0-9]+]]:_(s16), [[UV13:%[0-9]+]]:_(s16), [[UV14:%[0-9]+]]:_(s16) = G_UNMERGE_VALUES [[SHUF]](<7 x s16>)
|
|
; GFX9-NEXT: [[DEF:%[0-9]+]]:_(s16) = G_IMPLICIT_DEF
|
|
; GFX9-NEXT: [[BUILD_VECTOR2:%[0-9]+]]:_(<8 x s16>) = G_BUILD_VECTOR [[UV8]](s16), [[UV9]](s16), [[UV10]](s16), [[UV11]](s16), [[UV12]](s16), [[UV13]](s16), [[UV14]](s16), [[DEF]](s16)
|
|
; GFX9-NEXT: [[UV15:%[0-9]+]]:_(<2 x s16>), [[UV16:%[0-9]+]]:_(<2 x s16>), [[UV17:%[0-9]+]]:_(<2 x s16>), [[UV18:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[BUILD_VECTOR2]](<8 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV15]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV16]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr2 = COPY [[UV17]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr3 = COPY [[UV18]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
|
|
%res = shufflevector <7 x bfloat> %arg0, <7 x bfloat> zeroinitializer, <7 x i32> <i32 3, i32 1, i32 2, i32 0, i32 4, i32 5, i32 6>
|
|
ret <7 x bfloat> %res
|
|
}
|
|
|
|
define <8 x bfloat> @v8bf16(<8 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v8bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
|
|
; GFX9-NEXT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<8 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>)
|
|
; GFX9-NEXT: [[C:%[0-9]+]]:_(s16) = G_FCONSTANT bfloat 0.000000e+00
|
|
; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<8 x s16>) = G_BUILD_VECTOR [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16), [[C]](s16)
|
|
; GFX9-NEXT: [[SHUF:%[0-9]+]]:_(<8 x s16>) = G_SHUFFLE_VECTOR [[CONCAT_VECTORS]](<8 x s16>), [[BUILD_VECTOR]], shufflemask(3, 1, 2, 0, 4, 5, 6, 7)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[SHUF]](<8 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV1]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr2 = COPY [[UV2]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr3 = COPY [[UV3]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3
|
|
%res = shufflevector <8 x bfloat> %arg0, <8 x bfloat> zeroinitializer, <8 x i32> <i32 3, i32 1, i32 2, i32 0, i32 4, i32 5, i32 6, i32 7>
|
|
ret <8 x bfloat> %res
|
|
}
|
|
|
|
define <16 x bfloat> @v16bf16(<16 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v16bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
|
|
; GFX9-NEXT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
|
|
; GFX9-NEXT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
|
|
; GFX9-NEXT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
|
|
; GFX9-NEXT: [[COPY6:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr6
|
|
; GFX9-NEXT: [[COPY7:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr7
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<16 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[COPY6]](<2 x s16>), [[COPY7]](<2 x s16>)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>), [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<16 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV1]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr2 = COPY [[UV2]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr3 = COPY [[UV3]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr4 = COPY [[UV4]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr5 = COPY [[UV5]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr6 = COPY [[UV6]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr7 = COPY [[UV7]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7
|
|
ret <16 x bfloat> %arg0
|
|
}
|
|
|
|
define <32 x bfloat> @v32bf16(<32 x bfloat> %arg0) {
|
|
; GFX9-LABEL: name: v32bf16
|
|
; GFX9: bb.1 (%ir-block.0):
|
|
; GFX9-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5, $vgpr6, $vgpr7, $vgpr8, $vgpr9, $vgpr10, $vgpr11, $vgpr12, $vgpr13, $vgpr14, $vgpr15
|
|
; GFX9-NEXT: {{ $}}
|
|
; GFX9-NEXT: [[COPY:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr0
|
|
; GFX9-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr1
|
|
; GFX9-NEXT: [[COPY2:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr2
|
|
; GFX9-NEXT: [[COPY3:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr3
|
|
; GFX9-NEXT: [[COPY4:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr4
|
|
; GFX9-NEXT: [[COPY5:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr5
|
|
; GFX9-NEXT: [[COPY6:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr6
|
|
; GFX9-NEXT: [[COPY7:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr7
|
|
; GFX9-NEXT: [[COPY8:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr8
|
|
; GFX9-NEXT: [[COPY9:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr9
|
|
; GFX9-NEXT: [[COPY10:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr10
|
|
; GFX9-NEXT: [[COPY11:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr11
|
|
; GFX9-NEXT: [[COPY12:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr12
|
|
; GFX9-NEXT: [[COPY13:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr13
|
|
; GFX9-NEXT: [[COPY14:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr14
|
|
; GFX9-NEXT: [[COPY15:%[0-9]+]]:_(<2 x s16>) = COPY $vgpr15
|
|
; GFX9-NEXT: [[CONCAT_VECTORS:%[0-9]+]]:_(<32 x s16>) = G_CONCAT_VECTORS [[COPY]](<2 x s16>), [[COPY1]](<2 x s16>), [[COPY2]](<2 x s16>), [[COPY3]](<2 x s16>), [[COPY4]](<2 x s16>), [[COPY5]](<2 x s16>), [[COPY6]](<2 x s16>), [[COPY7]](<2 x s16>), [[COPY8]](<2 x s16>), [[COPY9]](<2 x s16>), [[COPY10]](<2 x s16>), [[COPY11]](<2 x s16>), [[COPY12]](<2 x s16>), [[COPY13]](<2 x s16>), [[COPY14]](<2 x s16>), [[COPY15]](<2 x s16>)
|
|
; GFX9-NEXT: [[UV:%[0-9]+]]:_(<2 x s16>), [[UV1:%[0-9]+]]:_(<2 x s16>), [[UV2:%[0-9]+]]:_(<2 x s16>), [[UV3:%[0-9]+]]:_(<2 x s16>), [[UV4:%[0-9]+]]:_(<2 x s16>), [[UV5:%[0-9]+]]:_(<2 x s16>), [[UV6:%[0-9]+]]:_(<2 x s16>), [[UV7:%[0-9]+]]:_(<2 x s16>), [[UV8:%[0-9]+]]:_(<2 x s16>), [[UV9:%[0-9]+]]:_(<2 x s16>), [[UV10:%[0-9]+]]:_(<2 x s16>), [[UV11:%[0-9]+]]:_(<2 x s16>), [[UV12:%[0-9]+]]:_(<2 x s16>), [[UV13:%[0-9]+]]:_(<2 x s16>), [[UV14:%[0-9]+]]:_(<2 x s16>), [[UV15:%[0-9]+]]:_(<2 x s16>) = G_UNMERGE_VALUES [[CONCAT_VECTORS]](<32 x s16>)
|
|
; GFX9-NEXT: $vgpr0 = COPY [[UV]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr1 = COPY [[UV1]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr2 = COPY [[UV2]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr3 = COPY [[UV3]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr4 = COPY [[UV4]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr5 = COPY [[UV5]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr6 = COPY [[UV6]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr7 = COPY [[UV7]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr8 = COPY [[UV8]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr9 = COPY [[UV9]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr10 = COPY [[UV10]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr11 = COPY [[UV11]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr12 = COPY [[UV12]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr13 = COPY [[UV13]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr14 = COPY [[UV14]](<2 x s16>)
|
|
; GFX9-NEXT: $vgpr15 = COPY [[UV15]](<2 x s16>)
|
|
; GFX9-NEXT: SI_RETURN implicit $vgpr0, implicit $vgpr1, implicit $vgpr2, implicit $vgpr3, implicit $vgpr4, implicit $vgpr5, implicit $vgpr6, implicit $vgpr7, implicit $vgpr8, implicit $vgpr9, implicit $vgpr10, implicit $vgpr11, implicit $vgpr12, implicit $vgpr13, implicit $vgpr14, implicit $vgpr15
|
|
ret <32 x bfloat> %arg0
|
|
}
|