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
104 lines
4.9 KiB
LLVM
104 lines
4.9 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2 < %s | FileCheck %s
|
|
|
|
;
|
|
; LDNT1B, LDNT1W, LDNT1H, LDNT1D: base + 32-bit unscaled offsets, zero (uxtw)
|
|
; extended to 64 bits.
|
|
; e.g. ldnt1h { z0.s }, p0/z, [z0.s, x0]
|
|
;
|
|
|
|
; LDNT1B
|
|
define <vscale x 4 x i32> @gldnt1b_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: gldnt1b_s_uxtw:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ldnt1b { z0.s }, p0/z, [z0.s, x0]
|
|
; CHECK-NEXT: ret
|
|
%load = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %pg,
|
|
ptr %base,
|
|
<vscale x 4 x i32> %b)
|
|
%res = zext <vscale x 4 x i8> %load to <vscale x 4 x i32>
|
|
ret <vscale x 4 x i32> %res
|
|
}
|
|
|
|
; LDNT1H
|
|
define <vscale x 4 x i32> @gldnt1h_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: gldnt1h_s_uxtw:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ldnt1h { z0.s }, p0/z, [z0.s, x0]
|
|
; CHECK-NEXT: ret
|
|
%load = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %pg,
|
|
ptr %base,
|
|
<vscale x 4 x i32> %b)
|
|
%res = zext <vscale x 4 x i16> %load to <vscale x 4 x i32>
|
|
ret <vscale x 4 x i32> %res
|
|
}
|
|
|
|
; LDNT1W
|
|
define <vscale x 4 x i32> @gldnt1w_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: gldnt1w_s_uxtw:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ldnt1w { z0.s }, p0/z, [z0.s, x0]
|
|
; CHECK-NEXT: ret
|
|
%load = call <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1> %pg,
|
|
ptr %base,
|
|
<vscale x 4 x i32> %b)
|
|
ret <vscale x 4 x i32> %load
|
|
}
|
|
|
|
define <vscale x 4 x float> @gldnt1w_s_uxtw_float(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: gldnt1w_s_uxtw_float:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ldnt1w { z0.s }, p0/z, [z0.s, x0]
|
|
; CHECK-NEXT: ret
|
|
%load = call <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1> %pg,
|
|
ptr %base,
|
|
<vscale x 4 x i32> %b)
|
|
ret <vscale x 4 x float> %load
|
|
}
|
|
|
|
; LDNT1SB, LDNT1SW, LDNT1SH: base + 32-bit unscaled offsets, zero (uxtw)
|
|
; extended to 64 bits.
|
|
; e.g. ldnt1sh { z0.s }, p0/z, [z0.s, x0]
|
|
;
|
|
|
|
; LDNT1SB
|
|
define <vscale x 4 x i32> @gldnt1sb_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: gldnt1sb_s_uxtw:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ldnt1sb { z0.s }, p0/z, [z0.s, x0]
|
|
; CHECK-NEXT: ret
|
|
%load = call <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1> %pg,
|
|
ptr %base,
|
|
<vscale x 4 x i32> %b)
|
|
%res = sext <vscale x 4 x i8> %load to <vscale x 4 x i32>
|
|
ret <vscale x 4 x i32> %res
|
|
}
|
|
|
|
; LDNT1SH
|
|
define <vscale x 4 x i32> @gldnt1sh_s_uxtw(<vscale x 4 x i1> %pg, ptr %base, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: gldnt1sh_s_uxtw:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: ldnt1sh { z0.s }, p0/z, [z0.s, x0]
|
|
; CHECK-NEXT: ret
|
|
%load = call <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1> %pg,
|
|
ptr %base,
|
|
<vscale x 4 x i32> %b)
|
|
%res = sext <vscale x 4 x i16> %load to <vscale x 4 x i32>
|
|
ret <vscale x 4 x i32> %res
|
|
}
|
|
|
|
; LDNT1B/LDNT1SB
|
|
declare <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i8(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x i8> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4i8(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
|
|
; LDNT1H/LDNT1SH
|
|
declare <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4i16(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x i16> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i16(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
|
|
; LDNT1W/LDNT1SW
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4i32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4i32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
|
|
declare <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.sxtw.nxv4f32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.gather.uxtw.nxv4f32(<vscale x 4 x i1>, ptr, <vscale x 4 x i32>)
|