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
333 lines
15 KiB
LLVM
333 lines
15 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=riscv64 -mattr=+v,+zvl256b | FileCheck %s
|
|
|
|
; Test binary operator with vp.merge and vp.smax.
|
|
define <8 x i32> @vpmerge_vpadd(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpadd:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.add.v8i32(<8 x i32> %x, <8 x i32> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test glued node of merge should not be deleted.
|
|
define <8 x i32> @vpmerge_vpadd2(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpadd2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vmseq.vv v0, v9, v10
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.add.v8i32(<8 x i32> %x, <8 x i32> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%m = call <8 x i1> @llvm.vp.icmp.v8i32(<8 x i32> %x, <8 x i32> %y, metadata !"eq", <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test vp.merge have all-ones mask.
|
|
define <8 x i32> @vpmerge_vpadd3(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpadd3:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vadd.vv v9, v9, v10
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, ma
|
|
; CHECK-NEXT: vmv.v.v v8, v9
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.add.v8i32(<8 x i32> %x, <8 x i32> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> splat (i1 true), <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test float binary operator with vp.merge and vp.fadd.
|
|
define <8 x float> @vpmerge_vpfadd(<8 x float> %passthru, <8 x float> %x, <8 x float> %y, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpfadd:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vfadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x float> @llvm.vp.fadd.v8f32(<8 x float> %x, <8 x float> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x float> @llvm.vp.merge.v8f32(<8 x i1> %m, <8 x float> %a, <8 x float> %passthru, i32 %vl)
|
|
ret <8 x float> %b
|
|
}
|
|
|
|
; Test conversion by fptosi.
|
|
define <8 x i16> @vpmerge_vpfptosi(<8 x i16> %passthru, <8 x float> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpfptosi:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e16, mf2, tu, mu
|
|
; CHECK-NEXT: vfncvt.rtz.x.f.w v8, v9, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i16> @llvm.vp.fptosi.v8i16.v8f32(<8 x float> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i16> @llvm.vp.merge.v8i16(<8 x i1> %m, <8 x i16> %a, <8 x i16> %passthru, i32 %vl)
|
|
ret <8 x i16> %b
|
|
}
|
|
|
|
; Test conversion by sitofp.
|
|
define <8 x float> @vpmerge_vpsitofp(<8 x float> %passthru, <8 x i64> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpsitofp:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vfncvt.f.x.w v8, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x float> @llvm.vp.sitofp.v8f32.v8i64(<8 x i64> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x float> @llvm.vp.merge.v8f32(<8 x i1> %m, <8 x float> %a, <8 x float> %passthru, i32 %vl)
|
|
ret <8 x float> %b
|
|
}
|
|
|
|
; Test integer extension by vp.zext.
|
|
define <8 x i32> @vpmerge_vpzext(<8 x i32> %passthru, <8 x i8> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpzext:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vzext.vf4 v8, v9, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.zext.v8i32.v8i8(<8 x i8> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test integer truncation by vp.trunc.
|
|
define <8 x i32> @vpmerge_vptrunc(<8 x i32> %passthru, <8 x i64> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vptrunc:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vnsrl.wi v8, v10, 0, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.trunc.v8i32.v8i64(<8 x i64> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test integer extension by vp.fpext.
|
|
define <8 x double> @vpmerge_vpfpext(<8 x double> %passthru, <8 x float> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpfpext:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vfwcvt.f.f.v v8, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x double> @llvm.vp.fpext.v8f64.v8f32(<8 x float> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x double> @llvm.vp.merge.v8f64(<8 x i1> %m, <8 x double> %a, <8 x double> %passthru, i32 %vl)
|
|
ret <8 x double> %b
|
|
}
|
|
|
|
; Test integer truncation by vp.trunc.
|
|
define <8 x float> @vpmerge_vpfptrunc(<8 x float> %passthru, <8 x double> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpfptrunc:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vfncvt.f.f.w v8, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x float> @llvm.vp.fptrunc.v8f32.v8f64(<8 x double> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x float> @llvm.vp.merge.v8f32(<8 x i1> %m, <8 x float> %a, <8 x float> %passthru, i32 %vl)
|
|
ret <8 x float> %b
|
|
}
|
|
|
|
; Test load operation by vp.load.
|
|
|
|
define <8 x i32> @vpmerge_vpload(<8 x i32> %passthru, ptr %p, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpload:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a1, e32, m1, tu, mu
|
|
; CHECK-NEXT: vle32.v v8, (a0), v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.load.v8i32.p0(ptr %p, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test result have chain and glued node.
|
|
define <8 x i32> @vpmerge_vpload2(<8 x i32> %passthru, ptr %p, <8 x i32> %x, <8 x i32> %y, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_vpload2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vmseq.vv v0, v9, v10
|
|
; CHECK-NEXT: vsetvli zero, a1, e32, m1, tu, mu
|
|
; CHECK-NEXT: vle32.v v8, (a0), v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.load.v8i32.p0(ptr %p, <8 x i1> splat (i1 true), i32 %vl)
|
|
%m = call <8 x i1> @llvm.vp.icmp.v8i32(<8 x i32> %x, <8 x i32> %y, metadata !"eq", <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test binary operator with vp.select and vp.add.
|
|
define <8 x i32> @vpselect_vpadd(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpadd:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.add.v8i32(<8 x i32> %x, <8 x i32> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test glued node of select should not be deleted.
|
|
define <8 x i32> @vpselect_vpadd2(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpadd2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vmseq.vv v0, v9, v10
|
|
; CHECK-NEXT: vadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.add.v8i32(<8 x i32> %x, <8 x i32> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%m = call <8 x i1> @llvm.vp.icmp.v8i32(<8 x i32> %x, <8 x i32> %y, metadata !"eq", <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test vp.select have all-ones mask.
|
|
define <8 x i32> @vpselect_vpadd3(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpadd3:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vadd.vv v8, v9, v10
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.add.v8i32(<8 x i32> %x, <8 x i32> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> splat (i1 true), <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test float binary operator with vp.select and vp.fadd.
|
|
define <8 x float> @vpselect_vpfadd(<8 x float> %passthru, <8 x float> %x, <8 x float> %y, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpfadd:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vfadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x float> @llvm.vp.fadd.v8f32(<8 x float> %x, <8 x float> %y, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x float> @llvm.vp.select.v8f32(<8 x i1> %m, <8 x float> %a, <8 x float> %passthru, i32 %vl)
|
|
ret <8 x float> %b
|
|
}
|
|
|
|
; Test conversion by fptosi.
|
|
define <8 x i16> @vpselect_vpfptosi(<8 x i16> %passthru, <8 x float> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpfptosi:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e16, mf2, ta, mu
|
|
; CHECK-NEXT: vfncvt.rtz.x.f.w v8, v9, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i16> @llvm.vp.fptosi.v8i16.v8f32(<8 x float> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i16> @llvm.vp.select.v8i16(<8 x i1> %m, <8 x i16> %a, <8 x i16> %passthru, i32 %vl)
|
|
ret <8 x i16> %b
|
|
}
|
|
|
|
; Test conversion by sitofp.
|
|
define <8 x float> @vpselect_vpsitofp(<8 x float> %passthru, <8 x i64> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpsitofp:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vfncvt.f.x.w v8, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x float> @llvm.vp.sitofp.v8f32.v8i64(<8 x i64> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x float> @llvm.vp.select.v8f32(<8 x i1> %m, <8 x float> %a, <8 x float> %passthru, i32 %vl)
|
|
ret <8 x float> %b
|
|
}
|
|
|
|
; Test integer extension by vp.zext.
|
|
define <8 x i32> @vpselect_vpzext(<8 x i32> %passthru, <8 x i8> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpzext:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vzext.vf4 v8, v9, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.zext.v8i32.v8i8(<8 x i8> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test integer truncation by vp.trunc.
|
|
define <8 x i32> @vpselect_vptrunc(<8 x i32> %passthru, <8 x i64> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vptrunc:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vnsrl.wi v8, v10, 0, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.trunc.v8i32.v8i64(<8 x i64> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test integer extension by vp.fpext.
|
|
define <8 x double> @vpselect_vpfpext(<8 x double> %passthru, <8 x float> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpfpext:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vfwcvt.f.f.v v8, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x double> @llvm.vp.fpext.v8f64.v8f32(<8 x float> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x double> @llvm.vp.select.v8f64(<8 x i1> %m, <8 x double> %a, <8 x double> %passthru, i32 %vl)
|
|
ret <8 x double> %b
|
|
}
|
|
|
|
; Test integer truncation by vp.trunc.
|
|
define <8 x float> @vpselect_vpfptrunc(<8 x float> %passthru, <8 x double> %x, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpfptrunc:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, mu
|
|
; CHECK-NEXT: vfncvt.f.f.w v8, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x float> @llvm.vp.fptrunc.v8f32.v8f64(<8 x double> %x, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x float> @llvm.vp.select.v8f32(<8 x i1> %m, <8 x float> %a, <8 x float> %passthru, i32 %vl)
|
|
ret <8 x float> %b
|
|
}
|
|
|
|
; Test load operation by vp.load.
|
|
define <8 x i32> @vpselect_vpload(<8 x i32> %passthru, ptr %p, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpload:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a1, e32, m1, ta, ma
|
|
; CHECK-NEXT: vle32.v v9, (a0)
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vmerge.vvm v8, v8, v9, v0
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.load.v8i32.p0(ptr %p, <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Test result have chain and glued node.
|
|
define <8 x i32> @vpselect_vpload2(<8 x i32> %passthru, ptr %p, <8 x i32> %x, <8 x i32> %y, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpselect_vpload2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a1, e32, m1, ta, ma
|
|
; CHECK-NEXT: vle32.v v11, (a0)
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vmseq.vv v0, v9, v10
|
|
; CHECK-NEXT: vmerge.vvm v8, v8, v11, v0
|
|
; CHECK-NEXT: ret
|
|
%a = call <8 x i32> @llvm.vp.load.v8i32.p0(ptr %p, <8 x i1> splat (i1 true), i32 %vl)
|
|
%m = call <8 x i1> @llvm.vp.icmp.v8i32(<8 x i32> %x, <8 x i32> %y, metadata !"eq", <8 x i1> splat (i1 true), i32 %vl)
|
|
%b = call <8 x i32> @llvm.vp.select.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
define <8 x i32> @vpmerge_add(<8 x i32> %passthru, <8 x i32> %x, <8 x i32> %y, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_add:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu
|
|
; CHECK-NEXT: vadd.vv v8, v9, v10, v0.t
|
|
; CHECK-NEXT: ret
|
|
%a = add <8 x i32> %x, %y
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|
|
|
|
; Negative test: Shouldn't fold vmerge into the load because we may increase vl.
|
|
define <8 x i32> @vpmerge_load(<8 x i32> %passthru, ptr %p, <8 x i1> %m, i32 zeroext %vl) {
|
|
; CHECK-LABEL: vpmerge_load:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: vsetivli zero, 8, e32, m1, ta, ma
|
|
; CHECK-NEXT: vle32.v v9, (a0)
|
|
; CHECK-NEXT: vsetvli zero, a1, e32, m1, tu, ma
|
|
; CHECK-NEXT: vmerge.vvm v8, v8, v9, v0
|
|
; CHECK-NEXT: ret
|
|
%a = load <8 x i32>, ptr %p
|
|
%b = call <8 x i32> @llvm.vp.merge.v8i32(<8 x i1> %m, <8 x i32> %a, <8 x i32> %passthru, i32 %vl)
|
|
ret <8 x i32> %b
|
|
}
|