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
204 lines
14 KiB
LLVM
204 lines
14 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc < %s -mtriple=arm -mattr=+vfp4d16sp,+fullfp16,-bf16 -stop-after=finalize-isel | FileCheck %s --check-prefixes=CHECK-NOBF16
|
|
|
|
; Check that the output instructions have the same fast math flags as the input
|
|
; fadd, even when bf16 is legalized to f32.
|
|
; FIXME: We should also test with +bf16, but it currently fails in instruction
|
|
; selection.
|
|
|
|
define bfloat @normal_fadd(bfloat %x, bfloat %y) {
|
|
; CHECK-NOBF16-LABEL: name: normal_fadd
|
|
; CHECK-NOBF16: bb.0.entry:
|
|
; CHECK-NOBF16-NEXT: liveins: $r0, $r1
|
|
; CHECK-NOBF16-NEXT: {{ $}}
|
|
; CHECK-NOBF16-NEXT: [[COPY:%[0-9]+]]:gpr = COPY $r1
|
|
; CHECK-NOBF16-NEXT: [[COPY1:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi:%[0-9]+]]:gpr = MOVsi [[COPY]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR:%[0-9]+]]:spr = VMOVSR killed [[MOVsi]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi1:%[0-9]+]]:gpr = MOVsi [[COPY1]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR1:%[0-9]+]]:spr = VMOVSR killed [[MOVsi1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS:%[0-9]+]]:spr = nofpexcept VADDS killed [[VMOVSR1]], killed [[VMOVSR]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS:%[0-9]+]]:gpr = VMOVRS killed [[VADDS]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY2:%[0-9]+]]:rgpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[VMOVHR:%[0-9]+]]:hpr = VMOVHR [[COPY2]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVRH:%[0-9]+]]:rgpr = VMOVRH killed [[VMOVHR]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRH]]
|
|
; CHECK-NOBF16-NEXT: MOVPCLR 14 /* CC::al */, $noreg, implicit $r0
|
|
entry:
|
|
%add = fadd bfloat %x, %y
|
|
ret bfloat %add
|
|
}
|
|
|
|
define bfloat @fast_fadd(bfloat %x, bfloat %y) {
|
|
; CHECK-NOBF16-LABEL: name: fast_fadd
|
|
; CHECK-NOBF16: bb.0.entry:
|
|
; CHECK-NOBF16-NEXT: liveins: $r0, $r1
|
|
; CHECK-NOBF16-NEXT: {{ $}}
|
|
; CHECK-NOBF16-NEXT: [[COPY:%[0-9]+]]:gpr = COPY $r1
|
|
; CHECK-NOBF16-NEXT: [[COPY1:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi:%[0-9]+]]:gpr = MOVsi [[COPY]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR:%[0-9]+]]:spr = VMOVSR killed [[MOVsi]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi1:%[0-9]+]]:gpr = MOVsi [[COPY1]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR1:%[0-9]+]]:spr = VMOVSR killed [[MOVsi1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS:%[0-9]+]]:spr = nnan ninf nsz arcp contract afn reassoc nofpexcept VADDS killed [[VMOVSR1]], killed [[VMOVSR]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS:%[0-9]+]]:gpr = VMOVRS killed [[VADDS]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY2:%[0-9]+]]:rgpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[VMOVHR:%[0-9]+]]:hpr = VMOVHR [[COPY2]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVRH:%[0-9]+]]:rgpr = VMOVRH killed [[VMOVHR]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRH]]
|
|
; CHECK-NOBF16-NEXT: MOVPCLR 14 /* CC::al */, $noreg, implicit $r0
|
|
entry:
|
|
%add = fadd fast bfloat %x, %y
|
|
ret bfloat %add
|
|
}
|
|
|
|
define bfloat @ninf_fadd(bfloat %x, bfloat %y) {
|
|
; CHECK-NOBF16-LABEL: name: ninf_fadd
|
|
; CHECK-NOBF16: bb.0.entry:
|
|
; CHECK-NOBF16-NEXT: liveins: $r0, $r1
|
|
; CHECK-NOBF16-NEXT: {{ $}}
|
|
; CHECK-NOBF16-NEXT: [[COPY:%[0-9]+]]:gpr = COPY $r1
|
|
; CHECK-NOBF16-NEXT: [[COPY1:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi:%[0-9]+]]:gpr = MOVsi [[COPY]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR:%[0-9]+]]:spr = VMOVSR killed [[MOVsi]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi1:%[0-9]+]]:gpr = MOVsi [[COPY1]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR1:%[0-9]+]]:spr = VMOVSR killed [[MOVsi1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS:%[0-9]+]]:spr = ninf nofpexcept VADDS killed [[VMOVSR1]], killed [[VMOVSR]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS:%[0-9]+]]:gpr = VMOVRS killed [[VADDS]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY2:%[0-9]+]]:rgpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[VMOVHR:%[0-9]+]]:hpr = VMOVHR [[COPY2]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVRH:%[0-9]+]]:rgpr = VMOVRH killed [[VMOVHR]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRH]]
|
|
; CHECK-NOBF16-NEXT: MOVPCLR 14 /* CC::al */, $noreg, implicit $r0
|
|
entry:
|
|
%add = fadd ninf bfloat %x, %y
|
|
ret bfloat %add
|
|
}
|
|
|
|
; Check that when we have the right fast math flags the converts in between the
|
|
; two fadds are removed.
|
|
|
|
define bfloat @normal_fadd_sequence(bfloat %x, bfloat %y, bfloat %z) {
|
|
; CHECK-NOBF16-LABEL: name: normal_fadd_sequence
|
|
; CHECK-NOBF16: bb.0.entry:
|
|
; CHECK-NOBF16-NEXT: liveins: $r0, $r1, $r2
|
|
; CHECK-NOBF16-NEXT: {{ $}}
|
|
; CHECK-NOBF16-NEXT: [[COPY:%[0-9]+]]:gpr = COPY $r2
|
|
; CHECK-NOBF16-NEXT: [[COPY1:%[0-9]+]]:gpr = COPY $r1
|
|
; CHECK-NOBF16-NEXT: [[COPY2:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi:%[0-9]+]]:gpr = MOVsi [[COPY1]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR:%[0-9]+]]:spr = VMOVSR killed [[MOVsi]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi1:%[0-9]+]]:gpr = MOVsi [[COPY2]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR1:%[0-9]+]]:spr = VMOVSR killed [[MOVsi1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS:%[0-9]+]]:spr = nofpexcept VADDS killed [[VMOVSR1]], killed [[VMOVSR]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS:%[0-9]+]]:gpr = VMOVRS killed [[VADDS]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY3:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi2:%[0-9]+]]:gpr = MOVsi [[COPY]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR2:%[0-9]+]]:spr = VMOVSR killed [[MOVsi2]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi3:%[0-9]+]]:gpr = MOVsi [[COPY3]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR3:%[0-9]+]]:spr = VMOVSR killed [[MOVsi3]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS1:%[0-9]+]]:spr = nofpexcept VADDS killed [[VMOVSR3]], killed [[VMOVSR2]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS1:%[0-9]+]]:gpr = VMOVRS killed [[VADDS1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS1]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY4:%[0-9]+]]:rgpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[VMOVHR:%[0-9]+]]:hpr = VMOVHR [[COPY4]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVRH:%[0-9]+]]:rgpr = VMOVRH killed [[VMOVHR]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRH]]
|
|
; CHECK-NOBF16-NEXT: MOVPCLR 14 /* CC::al */, $noreg, implicit $r0
|
|
entry:
|
|
%add1 = fadd bfloat %x, %y
|
|
%add2 = fadd bfloat %add1, %z
|
|
ret bfloat %add2
|
|
}
|
|
|
|
define bfloat @nnan_ninf_contract_fadd_sequence(bfloat %x, bfloat %y, bfloat %z) {
|
|
; CHECK-NOBF16-LABEL: name: nnan_ninf_contract_fadd_sequence
|
|
; CHECK-NOBF16: bb.0.entry:
|
|
; CHECK-NOBF16-NEXT: liveins: $r0, $r1, $r2
|
|
; CHECK-NOBF16-NEXT: {{ $}}
|
|
; CHECK-NOBF16-NEXT: [[COPY:%[0-9]+]]:gpr = COPY $r2
|
|
; CHECK-NOBF16-NEXT: [[COPY1:%[0-9]+]]:gpr = COPY $r1
|
|
; CHECK-NOBF16-NEXT: [[COPY2:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi:%[0-9]+]]:gpr = MOVsi [[COPY1]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR:%[0-9]+]]:spr = VMOVSR killed [[MOVsi]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi1:%[0-9]+]]:gpr = MOVsi [[COPY2]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR1:%[0-9]+]]:spr = VMOVSR killed [[MOVsi1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS:%[0-9]+]]:spr = nnan ninf contract nofpexcept VADDS killed [[VMOVSR1]], killed [[VMOVSR]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[MOVsi2:%[0-9]+]]:gpr = MOVsi [[COPY]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR2:%[0-9]+]]:spr = VMOVSR killed [[MOVsi2]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS1:%[0-9]+]]:spr = nnan ninf contract nofpexcept VADDS killed [[VADDS]], killed [[VMOVSR2]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS:%[0-9]+]]:gpr = VMOVRS killed [[VADDS1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY3:%[0-9]+]]:rgpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[VMOVHR:%[0-9]+]]:hpr = VMOVHR [[COPY3]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVRH:%[0-9]+]]:rgpr = VMOVRH killed [[VMOVHR]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRH]]
|
|
; CHECK-NOBF16-NEXT: MOVPCLR 14 /* CC::al */, $noreg, implicit $r0
|
|
entry:
|
|
%add1 = fadd nnan ninf contract bfloat %x, %y
|
|
%add2 = fadd nnan ninf contract bfloat %add1, %z
|
|
ret bfloat %add2
|
|
}
|
|
|
|
define bfloat @ninf_fadd_sequence(bfloat %x, bfloat %y, bfloat %z) {
|
|
; CHECK-NOBF16-LABEL: name: ninf_fadd_sequence
|
|
; CHECK-NOBF16: bb.0.entry:
|
|
; CHECK-NOBF16-NEXT: liveins: $r0, $r1, $r2
|
|
; CHECK-NOBF16-NEXT: {{ $}}
|
|
; CHECK-NOBF16-NEXT: [[COPY:%[0-9]+]]:gpr = COPY $r2
|
|
; CHECK-NOBF16-NEXT: [[COPY1:%[0-9]+]]:gpr = COPY $r1
|
|
; CHECK-NOBF16-NEXT: [[COPY2:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi:%[0-9]+]]:gpr = MOVsi [[COPY1]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR:%[0-9]+]]:spr = VMOVSR killed [[MOVsi]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi1:%[0-9]+]]:gpr = MOVsi [[COPY2]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR1:%[0-9]+]]:spr = VMOVSR killed [[MOVsi1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS:%[0-9]+]]:spr = ninf nofpexcept VADDS killed [[VMOVSR1]], killed [[VMOVSR]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS:%[0-9]+]]:gpr = VMOVRS killed [[VADDS]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY3:%[0-9]+]]:gpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[MOVsi2:%[0-9]+]]:gpr = MOVsi [[COPY]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR2:%[0-9]+]]:spr = VMOVSR killed [[MOVsi2]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[MOVsi3:%[0-9]+]]:gpr = MOVsi [[COPY3]], 130, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVSR3:%[0-9]+]]:spr = VMOVSR killed [[MOVsi3]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VADDS1:%[0-9]+]]:spr = ninf nofpexcept VADDS killed [[VMOVSR3]], killed [[VMOVSR2]], 14 /* CC::al */, $noreg, implicit $fpscr
|
|
; CHECK-NOBF16-NEXT: [[VMOVRS1:%[0-9]+]]:gpr = VMOVRS killed [[VADDS1]], 14 /* CC::al */, $noreg
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKDOWN 0, 0, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRS1]]
|
|
; CHECK-NOBF16-NEXT: BL &__truncsfbf2, csr_aapcs, implicit-def dead $lr, implicit $sp, implicit $r0, implicit-def $sp, implicit-def $r0
|
|
; CHECK-NOBF16-NEXT: ADJCALLSTACKUP 0, -1, 14 /* CC::al */, $noreg, implicit-def dead $sp, implicit $sp
|
|
; CHECK-NOBF16-NEXT: [[COPY4:%[0-9]+]]:rgpr = COPY $r0
|
|
; CHECK-NOBF16-NEXT: [[VMOVHR:%[0-9]+]]:hpr = VMOVHR [[COPY4]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: [[VMOVRH:%[0-9]+]]:rgpr = VMOVRH killed [[VMOVHR]], 14, $noreg
|
|
; CHECK-NOBF16-NEXT: $r0 = COPY [[VMOVRH]]
|
|
; CHECK-NOBF16-NEXT: MOVPCLR 14 /* CC::al */, $noreg, implicit $r0
|
|
entry:
|
|
%add1 = fadd ninf bfloat %x, %y
|
|
%add2 = fadd ninf bfloat %add1, %z
|
|
ret bfloat %add2
|
|
}
|