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
353 lines
13 KiB
LLVM
353 lines
13 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=riscv32 -mattr=+d,+zfbfmin,+zvfbfmin,+v -target-abi=ilp32d \
|
|
; RUN: -verify-machineinstrs < %s | FileCheck %s
|
|
; RUN: llc -mtriple=riscv64 -mattr=+d,+zfbfmin,+zvfbfmin,+v -target-abi=lp64d \
|
|
; RUN: -verify-machineinstrs < %s | FileCheck %s
|
|
|
|
define bfloat @vreduce_fmin_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fmin_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vfredmin.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call bfloat @llvm.vector.reduce.fmin.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fmax_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fmax_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vfredmax.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call bfloat @llvm.vector.reduce.fmax.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fmin_nnan_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fmin_nnan_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vfredmin.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call nnan bfloat @llvm.vector.reduce.fmin.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fmax_nnan_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fmax_nnan_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vfredmax.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call nnan bfloat @llvm.vector.reduce.fmax.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fminimum_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fminimum_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vmfne.vv v8, v10, v10
|
|
; CHECK-NEXT: vcpop.m a0, v8
|
|
; CHECK-NEXT: beqz a0, .LBB4_2
|
|
; CHECK-NEXT: # %bb.1:
|
|
; CHECK-NEXT: lui a0, 523264
|
|
; CHECK-NEXT: fmv.w.x fa5, a0
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
; CHECK-NEXT: .LBB4_2:
|
|
; CHECK-NEXT: vfredmin.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call bfloat @llvm.vector.reduce.fminimum.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fmaximum_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fmaximum_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vmfne.vv v8, v10, v10
|
|
; CHECK-NEXT: vcpop.m a0, v8
|
|
; CHECK-NEXT: beqz a0, .LBB5_2
|
|
; CHECK-NEXT: # %bb.1:
|
|
; CHECK-NEXT: lui a0, 523264
|
|
; CHECK-NEXT: fmv.w.x fa5, a0
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
; CHECK-NEXT: .LBB5_2:
|
|
; CHECK-NEXT: vfredmax.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call bfloat @llvm.vector.reduce.fmaximum.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fminimum_nnan_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fminimum_nnan_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vfredmin.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call nnan bfloat @llvm.vector.reduce.fminimum.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
define bfloat @vreduce_fmaximum_nnan_nxv4bf16(<vscale x 4 x bfloat> %val) {
|
|
; CHECK-LABEL: vreduce_fmaximum_nnan_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vfredmax.vs v8, v10, v10
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%s = call nnan bfloat @llvm.vector.reduce.fmaximum.nxv4bf16(<vscale x 4 x bfloat> %val)
|
|
ret bfloat %s
|
|
}
|
|
|
|
|
|
define bfloat @vreduce_fadd_nxv1bf16(<vscale x 1 x bfloat> %v, bfloat %s) {
|
|
; CHECK-LABEL: vreduce_fadd_nxv1bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, mf4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
|
|
; CHECK-NEXT: lui a0, 524288
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, mf2, ta, ma
|
|
; CHECK-NEXT: vmv.s.x v8, a0
|
|
; CHECK-NEXT: vfredusum.vs v8, v9, v8
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa4, fa0
|
|
; CHECK-NEXT: fadd.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call reassoc bfloat @llvm.vector.reduce.fadd.nxv1bf16(bfloat %s, <vscale x 1 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fadd_nxv2bf16(<vscale x 2 x bfloat> %v, bfloat %s) {
|
|
; CHECK-LABEL: vreduce_fadd_nxv2bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: lui a0, 524288
|
|
; CHECK-NEXT: vsetvli a1, zero, e16, mf2, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v9, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m1, ta, ma
|
|
; CHECK-NEXT: vmv.s.x v8, a0
|
|
; CHECK-NEXT: vfredusum.vs v8, v9, v8
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa4, fa0
|
|
; CHECK-NEXT: fadd.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call reassoc bfloat @llvm.vector.reduce.fadd.nxv2bf16(bfloat %s, <vscale x 2 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fadd_nxv4bf16(<vscale x 4 x bfloat> %v, bfloat %s) {
|
|
; CHECK-LABEL: vreduce_fadd_nxv4bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: lui a0, 524288
|
|
; CHECK-NEXT: vsetvli a1, zero, e16, m1, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v10, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m2, ta, ma
|
|
; CHECK-NEXT: vmv.s.x v8, a0
|
|
; CHECK-NEXT: vfredusum.vs v8, v10, v8
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa4, fa0
|
|
; CHECK-NEXT: fadd.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call reassoc bfloat @llvm.vector.reduce.fadd.nxv4bf16(bfloat %s, <vscale x 4 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fmin_nxv32bf16(<vscale x 32 x bfloat> %v) {
|
|
; CHECK-LABEL: vreduce_fmin_nxv32bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vfredmin.vs v8, v16, v16
|
|
; CHECK-NEXT: vfredmin.vs v9, v24, v24
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: vfmv.f.s fa4, v9
|
|
; CHECK-NEXT: fmin.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call bfloat @llvm.vector.reduce.fmin.nxv32bf16(<vscale x 32 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fmax_nxv32bf16(<vscale x 32 x bfloat> %v) {
|
|
; CHECK-LABEL: vreduce_fmax_nxv32bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vfredmax.vs v8, v16, v16
|
|
; CHECK-NEXT: vfredmax.vs v9, v24, v24
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: vfmv.f.s fa4, v9
|
|
; CHECK-NEXT: fmax.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call bfloat @llvm.vector.reduce.fmax.nxv32bf16(<vscale x 32 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fminimum_nxv32bf16(<vscale x 32 x bfloat> %v) {
|
|
; CHECK-LABEL: vreduce_fminimum_nxv32bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vmfne.vv v12, v16, v16
|
|
; CHECK-NEXT: vcpop.m a0, v12
|
|
; CHECK-NEXT: lui a1, 523264
|
|
; CHECK-NEXT: fmv.w.x fa5, a1
|
|
; CHECK-NEXT: beqz a0, .LBB13_2
|
|
; CHECK-NEXT: # %bb.1:
|
|
; CHECK-NEXT: fmv.s fa4, fa5
|
|
; CHECK-NEXT: j .LBB13_3
|
|
; CHECK-NEXT: .LBB13_2:
|
|
; CHECK-NEXT: vfredmin.vs v12, v16, v16
|
|
; CHECK-NEXT: vfmv.f.s fa4, v12
|
|
; CHECK-NEXT: .LBB13_3:
|
|
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vmfne.vv v8, v16, v16
|
|
; CHECK-NEXT: vcpop.m a0, v8
|
|
; CHECK-NEXT: bnez a0, .LBB13_5
|
|
; CHECK-NEXT: # %bb.4:
|
|
; CHECK-NEXT: vfredmin.vs v8, v16, v16
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: .LBB13_5:
|
|
; CHECK-NEXT: feq.s a0, fa5, fa5
|
|
; CHECK-NEXT: fmv.s fa3, fa4
|
|
; CHECK-NEXT: bnez a0, .LBB13_7
|
|
; CHECK-NEXT: # %bb.6:
|
|
; CHECK-NEXT: fmv.s fa3, fa5
|
|
; CHECK-NEXT: .LBB13_7:
|
|
; CHECK-NEXT: feq.s a0, fa4, fa4
|
|
; CHECK-NEXT: bnez a0, .LBB13_9
|
|
; CHECK-NEXT: # %bb.8:
|
|
; CHECK-NEXT: fmv.s fa5, fa4
|
|
; CHECK-NEXT: .LBB13_9:
|
|
; CHECK-NEXT: fmin.s fa5, fa5, fa3
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call bfloat @llvm.vector.reduce.fminimum.nxv32bf16(<vscale x 32 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fmaximum_nxv32bf16(<vscale x 32 x bfloat> %v) {
|
|
; CHECK-LABEL: vreduce_fmaximum_nxv32bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli a0, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vmfne.vv v12, v16, v16
|
|
; CHECK-NEXT: vcpop.m a0, v12
|
|
; CHECK-NEXT: lui a1, 523264
|
|
; CHECK-NEXT: fmv.w.x fa5, a1
|
|
; CHECK-NEXT: beqz a0, .LBB14_2
|
|
; CHECK-NEXT: # %bb.1:
|
|
; CHECK-NEXT: fmv.s fa4, fa5
|
|
; CHECK-NEXT: j .LBB14_3
|
|
; CHECK-NEXT: .LBB14_2:
|
|
; CHECK-NEXT: vfredmax.vs v12, v16, v16
|
|
; CHECK-NEXT: vfmv.f.s fa4, v12
|
|
; CHECK-NEXT: .LBB14_3:
|
|
; CHECK-NEXT: vsetvli zero, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vmfne.vv v8, v16, v16
|
|
; CHECK-NEXT: vcpop.m a0, v8
|
|
; CHECK-NEXT: bnez a0, .LBB14_5
|
|
; CHECK-NEXT: # %bb.4:
|
|
; CHECK-NEXT: vfredmax.vs v8, v16, v16
|
|
; CHECK-NEXT: vfmv.f.s fa5, v8
|
|
; CHECK-NEXT: .LBB14_5:
|
|
; CHECK-NEXT: feq.s a0, fa5, fa5
|
|
; CHECK-NEXT: fmv.s fa3, fa4
|
|
; CHECK-NEXT: bnez a0, .LBB14_7
|
|
; CHECK-NEXT: # %bb.6:
|
|
; CHECK-NEXT: fmv.s fa3, fa5
|
|
; CHECK-NEXT: .LBB14_7:
|
|
; CHECK-NEXT: feq.s a0, fa4, fa4
|
|
; CHECK-NEXT: bnez a0, .LBB14_9
|
|
; CHECK-NEXT: # %bb.8:
|
|
; CHECK-NEXT: fmv.s fa5, fa4
|
|
; CHECK-NEXT: .LBB14_9:
|
|
; CHECK-NEXT: fmax.s fa5, fa5, fa3
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call bfloat @llvm.vector.reduce.fmaximum.nxv32bf16(<vscale x 32 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|
|
|
|
define bfloat @vreduce_fadd_nxv32bf16(<vscale x 32 x bfloat> %v, bfloat %s) {
|
|
; CHECK-LABEL: vreduce_fadd_nxv32bf16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: lui a0, 524288
|
|
; CHECK-NEXT: vsetvli a1, zero, e16, m4, ta, ma
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v16, v12
|
|
; CHECK-NEXT: vfwcvtbf16.f.f.v v24, v8
|
|
; CHECK-NEXT: vsetvli zero, zero, e32, m8, ta, ma
|
|
; CHECK-NEXT: vmv.s.x v8, a0
|
|
; CHECK-NEXT: vfredusum.vs v9, v16, v8
|
|
; CHECK-NEXT: vfredusum.vs v8, v24, v8
|
|
; CHECK-NEXT: vfmv.f.s fa5, v9
|
|
; CHECK-NEXT: vfmv.f.s fa4, v8
|
|
; CHECK-NEXT: fcvt.bf16.s fa5, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa4, fa4
|
|
; CHECK-NEXT: fcvt.s.bf16 fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa4, fa4
|
|
; CHECK-NEXT: fadd.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa5, fa5
|
|
; CHECK-NEXT: fcvt.s.bf16 fa4, fa0
|
|
; CHECK-NEXT: fadd.s fa5, fa4, fa5
|
|
; CHECK-NEXT: fcvt.bf16.s fa0, fa5
|
|
; CHECK-NEXT: ret
|
|
%red = call reassoc bfloat @llvm.vector.reduce.fadd.nxv32bf16(bfloat %s, <vscale x 32 x bfloat> %v)
|
|
ret bfloat %red
|
|
}
|