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
224 lines
8.1 KiB
LLVM
224 lines
8.1 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
|
|
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve,+sme < %s | FileCheck %s
|
|
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme -force-streaming < %s | FileCheck %s
|
|
|
|
;
|
|
; SQXTNB
|
|
;
|
|
|
|
define <vscale x 16 x i8> @sqxtnb_h(<vscale x 8 x i16> %a) {
|
|
; CHECK-LABEL: sqxtnb_h:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtnb z0.b, z0.h
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnb.nxv8i16(<vscale x 8 x i16> %a)
|
|
ret <vscale x 16 x i8> %out
|
|
}
|
|
|
|
define <vscale x 8 x i16> @sqxtnb_s(<vscale x 4 x i32> %a) {
|
|
; CHECK-LABEL: sqxtnb_s:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtnb z0.h, z0.s
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnb.nxv4i32(<vscale x 4 x i32> %a)
|
|
ret <vscale x 8 x i16> %out
|
|
}
|
|
|
|
define <vscale x 4 x i32> @sqxtnb_d(<vscale x 2 x i64> %a) {
|
|
; CHECK-LABEL: sqxtnb_d:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtnb z0.s, z0.d
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnb.nxv2i64(<vscale x 2 x i64> %a)
|
|
ret <vscale x 4 x i32> %out
|
|
}
|
|
|
|
;
|
|
; UQXTNB
|
|
;
|
|
|
|
define <vscale x 16 x i8> @uqxtnb_h(<vscale x 8 x i16> %a) {
|
|
; CHECK-LABEL: uqxtnb_h:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: uqxtnb z0.b, z0.h
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnb.nxv8i16(<vscale x 8 x i16> %a)
|
|
ret <vscale x 16 x i8> %out
|
|
}
|
|
|
|
define <vscale x 8 x i16> @uqxtnb_s(<vscale x 4 x i32> %a) {
|
|
; CHECK-LABEL: uqxtnb_s:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: uqxtnb z0.h, z0.s
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnb.nxv4i32(<vscale x 4 x i32> %a)
|
|
ret <vscale x 8 x i16> %out
|
|
}
|
|
|
|
define <vscale x 4 x i32> @uqxtnb_d(<vscale x 2 x i64> %a) {
|
|
; CHECK-LABEL: uqxtnb_d:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: uqxtnb z0.s, z0.d
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnb.nxv2i64(<vscale x 2 x i64> %a)
|
|
ret <vscale x 4 x i32> %out
|
|
}
|
|
|
|
;
|
|
; SQXTUNB
|
|
;
|
|
|
|
define <vscale x 16 x i8> @sqxtunb_h(<vscale x 8 x i16> %a) {
|
|
; CHECK-LABEL: sqxtunb_h:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtunb z0.b, z0.h
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunb.nxv8i16(<vscale x 8 x i16> %a)
|
|
ret <vscale x 16 x i8> %out
|
|
}
|
|
|
|
define <vscale x 8 x i16> @sqxtunb_s(<vscale x 4 x i32> %a) {
|
|
; CHECK-LABEL: sqxtunb_s:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtunb z0.h, z0.s
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunb.nxv4i32(<vscale x 4 x i32> %a)
|
|
ret <vscale x 8 x i16> %out
|
|
}
|
|
|
|
define <vscale x 4 x i32> @sqxtunb_d(<vscale x 2 x i64> %a) {
|
|
; CHECK-LABEL: sqxtunb_d:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtunb z0.s, z0.d
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunb.nxv2i64(<vscale x 2 x i64> %a)
|
|
ret <vscale x 4 x i32> %out
|
|
}
|
|
|
|
;
|
|
; SQXTNT
|
|
;
|
|
|
|
define <vscale x 16 x i8> @sqxtnt_h(<vscale x 16 x i8> %a, <vscale x 8 x i16> %b) {
|
|
; CHECK-LABEL: sqxtnt_h:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtnt z0.b, z1.h
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnt.nxv8i16(<vscale x 16 x i8> %a,
|
|
<vscale x 8 x i16> %b)
|
|
ret <vscale x 16 x i8> %out
|
|
}
|
|
|
|
define <vscale x 8 x i16> @sqxtnt_s(<vscale x 8 x i16> %a, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: sqxtnt_s:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtnt z0.h, z1.s
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnt.nxv4i32(<vscale x 8 x i16> %a,
|
|
<vscale x 4 x i32> %b)
|
|
ret <vscale x 8 x i16> %out
|
|
}
|
|
|
|
define <vscale x 4 x i32> @sqxtnt_d(<vscale x 4 x i32> %a, <vscale x 2 x i64> %b) {
|
|
; CHECK-LABEL: sqxtnt_d:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtnt z0.s, z1.d
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnt.nxv2i64(<vscale x 4 x i32> %a,
|
|
<vscale x 2 x i64> %b)
|
|
ret <vscale x 4 x i32> %out
|
|
}
|
|
|
|
;
|
|
; UQXTNT
|
|
;
|
|
|
|
define <vscale x 16 x i8> @uqxtnt_h(<vscale x 16 x i8> %a, <vscale x 8 x i16> %b) {
|
|
; CHECK-LABEL: uqxtnt_h:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: uqxtnt z0.b, z1.h
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnt.nxv8i16(<vscale x 16 x i8> %a,
|
|
<vscale x 8 x i16> %b)
|
|
ret <vscale x 16 x i8> %out
|
|
}
|
|
|
|
define <vscale x 8 x i16> @uqxtnt_s(<vscale x 8 x i16> %a, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: uqxtnt_s:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: uqxtnt z0.h, z1.s
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnt.nxv4i32(<vscale x 8 x i16> %a,
|
|
<vscale x 4 x i32> %b)
|
|
ret <vscale x 8 x i16> %out
|
|
}
|
|
|
|
define <vscale x 4 x i32> @uqxtnt_d(<vscale x 4 x i32> %a, <vscale x 2 x i64> %b) {
|
|
; CHECK-LABEL: uqxtnt_d:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: uqxtnt z0.s, z1.d
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnt.nxv2i64(<vscale x 4 x i32> %a,
|
|
<vscale x 2 x i64> %b)
|
|
ret <vscale x 4 x i32> %out
|
|
}
|
|
|
|
;
|
|
; SQXTUNT
|
|
;
|
|
|
|
define <vscale x 16 x i8> @sqxtunt_h(<vscale x 16 x i8> %a, <vscale x 8 x i16> %b) {
|
|
; CHECK-LABEL: sqxtunt_h:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtunt z0.b, z1.h
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunt.nxv8i16(<vscale x 16 x i8> %a,
|
|
<vscale x 8 x i16> %b)
|
|
ret <vscale x 16 x i8> %out
|
|
}
|
|
|
|
define <vscale x 8 x i16> @sqxtunt_s(<vscale x 8 x i16> %a, <vscale x 4 x i32> %b) {
|
|
; CHECK-LABEL: sqxtunt_s:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtunt z0.h, z1.s
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunt.nxv4i32(<vscale x 8 x i16> %a,
|
|
<vscale x 4 x i32> %b)
|
|
ret <vscale x 8 x i16> %out
|
|
}
|
|
|
|
define <vscale x 4 x i32> @sqxtunt_d(<vscale x 4 x i32> %a, <vscale x 2 x i64> %b) {
|
|
; CHECK-LABEL: sqxtunt_d:
|
|
; CHECK: // %bb.0:
|
|
; CHECK-NEXT: sqxtunt z0.s, z1.d
|
|
; CHECK-NEXT: ret
|
|
%out = call <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunt.nxv2i64(<vscale x 4 x i32> %a,
|
|
<vscale x 2 x i64> %b)
|
|
ret <vscale x 4 x i32> %out
|
|
}
|
|
|
|
declare <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnb.nxv8i16(<vscale x 8 x i16>)
|
|
declare <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnb.nxv4i32(<vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnb.nxv2i64(<vscale x 2 x i64>)
|
|
|
|
declare <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnb.nxv8i16(<vscale x 8 x i16>)
|
|
declare <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnb.nxv4i32(<vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnb.nxv2i64(<vscale x 2 x i64>)
|
|
|
|
declare <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunb.nxv8i16(<vscale x 8 x i16>)
|
|
declare <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunb.nxv4i32(<vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunb.nxv2i64(<vscale x 2 x i64>)
|
|
|
|
declare <vscale x 16 x i8> @llvm.aarch64.sve.sqxtnt.nxv8i16(<vscale x 16 x i8>, <vscale x 8 x i16>)
|
|
declare <vscale x 8 x i16> @llvm.aarch64.sve.sqxtnt.nxv4i32(<vscale x 8 x i16>, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.sqxtnt.nxv2i64(<vscale x 4 x i32>, <vscale x 2 x i64>)
|
|
|
|
declare <vscale x 16 x i8> @llvm.aarch64.sve.uqxtnt.nxv8i16(<vscale x 16 x i8>, <vscale x 8 x i16>)
|
|
declare <vscale x 8 x i16> @llvm.aarch64.sve.uqxtnt.nxv4i32(<vscale x 8 x i16>, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.uqxtnt.nxv2i64(<vscale x 4 x i32>, <vscale x 2 x i64>)
|
|
|
|
declare <vscale x 16 x i8> @llvm.aarch64.sve.sqxtunt.nxv8i16(<vscale x 16 x i8>, <vscale x 8 x i16>)
|
|
declare <vscale x 8 x i16> @llvm.aarch64.sve.sqxtunt.nxv4i32(<vscale x 8 x i16>, <vscale x 4 x i32>)
|
|
declare <vscale x 4 x i32> @llvm.aarch64.sve.sqxtunt.nxv2i64(<vscale x 4 x i32>, <vscale x 2 x i64>)
|