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
526 lines
21 KiB
LLVM
526 lines
21 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=riscv64 -mattr=+v -verify-machineinstrs \
|
|
; RUN: < %s | FileCheck %s --check-prefix=NOVLDEP
|
|
; RUN: llc -mtriple=riscv64 -mattr=+v,+vl-dependent-latency -verify-machineinstrs \
|
|
; RUN: < %s | FileCheck %s --check-prefix=VLDEP
|
|
|
|
define <2 x i1> @test_vp_splice_v2i1(<2 x i1> %va, <2 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v2i1:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v9, v9, 1
|
|
; NOVLDEP-NEXT: addi a0, a0, -1
|
|
; NOVLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v2i1:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -1
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v9, v9, 1
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <2 x i1> @llvm.experimental.vp.splice.v2i1(<2 x i1> %va, <2 x i1> %vb, i32 1, <2 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <2 x i1> %v
|
|
}
|
|
|
|
define <2 x i1> @test_vp_splice_v2i1_negative_offset(<2 x i1> %va, <2 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v2i1_negative_offset:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -1
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; NOVLDEP-NEXT: vslideup.vi v9, v8, 1
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v2i1_negative_offset:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -1
|
|
; VLDEP-NEXT: vsetivli zero, 1, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vslideup.vi v9, v8, 1
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <2 x i1> @llvm.experimental.vp.splice.v2i1(<2 x i1> %va, <2 x i1> %vb, i32 -1, <2 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <2 x i1> %v
|
|
}
|
|
|
|
define <2 x i1> @test_vp_splice_v2i1_masked(<2 x i1> %va, <2 x i1> %vb, <2 x i1> %mask, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v2i1_masked:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v10, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v11, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v10
|
|
; NOVLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -1
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v10, v10, 1, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, mf8, ta, mu
|
|
; NOVLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, mf8, ta, ma
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v2i1_masked:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v10, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v11, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v10
|
|
; VLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -1
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v10, v10, 1, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf8, ta, mu
|
|
; VLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, zero, e8, mf8, ta, ma
|
|
; VLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; VLDEP-NEXT: ret
|
|
%v = call <2 x i1> @llvm.experimental.vp.splice.v2i1(<2 x i1> %va, <2 x i1> %vb, i32 1, <2 x i1> %mask, i32 %evla, i32 %evlb)
|
|
ret <2 x i1> %v
|
|
}
|
|
|
|
define <4 x i1> @test_vp_splice_v4i1(<4 x i1> %va, <4 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v4i1:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v9, v9, 3
|
|
; NOVLDEP-NEXT: addi a0, a0, -3
|
|
; NOVLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v4i1:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -3
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v9, v9, 3
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <4 x i1> @llvm.experimental.vp.splice.v4i1(<4 x i1> %va, <4 x i1> %vb, i32 3, <4 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <4 x i1> %v
|
|
}
|
|
|
|
define <4 x i1> @test_vp_splice_v4i1_negative_offset(<4 x i1> %va, <4 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v4i1_negative_offset:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -3
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; NOVLDEP-NEXT: vslideup.vi v9, v8, 3
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v4i1_negative_offset:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -3
|
|
; VLDEP-NEXT: vsetivli zero, 3, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vslideup.vi v9, v8, 3
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <4 x i1> @llvm.experimental.vp.splice.v4i1(<4 x i1> %va, <4 x i1> %vb, i32 -3, <4 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <4 x i1> %v
|
|
}
|
|
|
|
define <4 x i1> @test_vp_splice_v4i1_masked(<4 x i1> %va, <4 x i1> %vb, <4 x i1> %mask, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v4i1_masked:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v10, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v11, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v10
|
|
; NOVLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -3
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v10, v10, 3, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, mf4, ta, mu
|
|
; NOVLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, mf4, ta, ma
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v4i1_masked:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v10, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v11, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v10
|
|
; VLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -3
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v10, v10, 3, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf4, ta, mu
|
|
; VLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, zero, e8, mf4, ta, ma
|
|
; VLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; VLDEP-NEXT: ret
|
|
%v = call <4 x i1> @llvm.experimental.vp.splice.v4i1(<4 x i1> %va, <4 x i1> %vb, i32 3, <4 x i1> %mask, i32 %evla, i32 %evlb)
|
|
ret <4 x i1> %v
|
|
}
|
|
|
|
define <8 x i1> @test_vp_splice_v8i1(<8 x i1> %va, <8 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v8i1:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v9, v9, 5
|
|
; NOVLDEP-NEXT: addi a0, a0, -5
|
|
; NOVLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v8i1:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -5
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v9, v9, 5
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <8 x i1> @llvm.experimental.vp.splice.v8i1(<8 x i1> %va, <8 x i1> %vb, i32 5, <8 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <8 x i1> %v
|
|
}
|
|
|
|
define <8 x i1> @test_vp_splice_v8i1_negative_offset(<8 x i1> %va, <8 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v8i1_negative_offset:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -5
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; NOVLDEP-NEXT: vslideup.vi v9, v8, 5
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v8i1_negative_offset:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -5
|
|
; VLDEP-NEXT: vsetivli zero, 5, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vslideup.vi v9, v8, 5
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <8 x i1> @llvm.experimental.vp.splice.v8i1(<8 x i1> %va, <8 x i1> %vb, i32 -5, <8 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <8 x i1> %v
|
|
}
|
|
|
|
define <8 x i1> @test_vp_splice_v8i1_masked(<8 x i1> %va, <8 x i1> %vb, <8 x i1> %mask, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v8i1_masked:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v10, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v11, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v10
|
|
; NOVLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -5
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v10, v10, 5, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, mf2, ta, mu
|
|
; NOVLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, mf2, ta, ma
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v8i1_masked:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v10, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v11, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v10
|
|
; VLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -5
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v10, v10, 5, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, mf2, ta, mu
|
|
; VLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, zero, e8, mf2, ta, ma
|
|
; VLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; VLDEP-NEXT: ret
|
|
%v = call <8 x i1> @llvm.experimental.vp.splice.v8i1(<8 x i1> %va, <8 x i1> %vb, i32 5, <8 x i1> %mask, i32 %evla, i32 %evlb)
|
|
ret <8 x i1> %v
|
|
}
|
|
|
|
define <16 x i1> @test_vp_splice_v16i1(<16 x i1> %va, <16 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v16i1:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v9, v9, 5
|
|
; NOVLDEP-NEXT: addi a0, a0, -5
|
|
; NOVLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v16i1:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -5
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v9, v9, 5
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vslideup.vx v9, v8, a0
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <16 x i1> @llvm.experimental.vp.splice.v16i1(<16 x i1> %va, <16 x i1> %vb, i32 5, <16 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <16 x i1> %v
|
|
}
|
|
|
|
define <16 x i1> @test_vp_splice_v16i1_negative_offset(<16 x i1> %va, <16 x i1> %vb, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v16i1_negative_offset:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v9, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v10, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -5
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; NOVLDEP-NEXT: vslideup.vi v9, v8, 5
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v16i1_negative_offset:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v9, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v10, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vmerge.vim v9, v10, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -5
|
|
; VLDEP-NEXT: vsetivli zero, 5, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vx v9, v9, a0
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vslideup.vi v9, v8, 5
|
|
; VLDEP-NEXT: vmsne.vi v0, v9, 0
|
|
; VLDEP-NEXT: ret
|
|
|
|
%v = call <16 x i1> @llvm.experimental.vp.splice.v16i1(<16 x i1> %va, <16 x i1> %vb, i32 -5, <16 x i1> splat (i1 1), i32 %evla, i32 %evlb)
|
|
ret <16 x i1> %v
|
|
}
|
|
|
|
define <16 x i1> @test_vp_splice_v16i1_masked(<16 x i1> %va, <16 x i1> %vb, <16 x i1> %mask, i32 zeroext %evla, i32 zeroext %evlb) {
|
|
; NOVLDEP-LABEL: test_vp_splice_v16i1_masked:
|
|
; NOVLDEP: # %bb.0:
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmv1r.v v10, v0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v8
|
|
; NOVLDEP-NEXT: vmv.v.i v8, 0
|
|
; NOVLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; NOVLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmv.v.i v11, 0
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v10
|
|
; NOVLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; NOVLDEP-NEXT: addi a0, a0, -5
|
|
; NOVLDEP-NEXT: vmv1r.v v0, v9
|
|
; NOVLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vslidedown.vi v10, v10, 5, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, m1, ta, mu
|
|
; NOVLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; NOVLDEP-NEXT: vsetvli zero, zero, e8, m1, ta, ma
|
|
; NOVLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; NOVLDEP-NEXT: ret
|
|
;
|
|
; VLDEP-LABEL: test_vp_splice_v16i1_masked:
|
|
; VLDEP: # %bb.0:
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmv1r.v v10, v0
|
|
; VLDEP-NEXT: vmv1r.v v0, v8
|
|
; VLDEP-NEXT: vmv.v.i v8, 0
|
|
; VLDEP-NEXT: vmerge.vim v8, v8, 1, v0
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmv.v.i v11, 0
|
|
; VLDEP-NEXT: vmv1r.v v0, v10
|
|
; VLDEP-NEXT: vmerge.vim v10, v11, 1, v0
|
|
; VLDEP-NEXT: addi a0, a0, -5
|
|
; VLDEP-NEXT: vmv1r.v v0, v9
|
|
; VLDEP-NEXT: vsetvli zero, a0, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vslidedown.vi v10, v10, 5, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, a1, e8, m1, ta, mu
|
|
; VLDEP-NEXT: vslideup.vx v10, v8, a0, v0.t
|
|
; VLDEP-NEXT: vsetvli zero, zero, e8, m1, ta, ma
|
|
; VLDEP-NEXT: vmsne.vi v0, v10, 0, v0.t
|
|
; VLDEP-NEXT: ret
|
|
%v = call <16 x i1> @llvm.experimental.vp.splice.v16i1(<16 x i1> %va, <16 x i1> %vb, i32 5, <16 x i1> %mask, i32 %evla, i32 %evlb)
|
|
ret <16 x i1> %v
|
|
}
|