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
206 lines
6.5 KiB
LLVM
206 lines
6.5 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
|
|
; RUN: -mattr=+vsx -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
|
|
; RUN: FileCheck %s
|
|
; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
|
|
; RUN: -mattr=+vsx -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
|
|
; RUN: FileCheck %s --check-prefix=CHECK-BE
|
|
|
|
define void @test8i8(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test8i8:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: vpkuhum v2, v2, v2
|
|
; CHECK-NEXT: xxswapd vs0, v2
|
|
; CHECK-NEXT: stfdx f0, 0, r3
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test8i8:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r5, r1, -16
|
|
; CHECK-BE-NEXT: vpkuhum v2, v2, v2
|
|
; CHECK-BE-NEXT: stxvd2x v2, 0, r5
|
|
; CHECK-BE-NEXT: ld r4, -16(r1)
|
|
; CHECK-BE-NEXT: std r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <8 x i16>, ptr %SrcPtr, align 16
|
|
%1 = trunc <8 x i16> %0 to <8 x i8>
|
|
store <8 x i8> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test4i8(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test4i8:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: vpkuhum v2, v2, v2
|
|
; CHECK-NEXT: xxsldwi vs0, v2, v2, 2
|
|
; CHECK-NEXT: stfiwx f0, 0, r3
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test4i8:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r5, r1, -16
|
|
; CHECK-BE-NEXT: vpkuhum v2, v2, v2
|
|
; CHECK-BE-NEXT: stxvw4x v2, 0, r5
|
|
; CHECK-BE-NEXT: lwz r4, -16(r1)
|
|
; CHECK-BE-NEXT: stw r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <4 x i16>, ptr %SrcPtr, align 16
|
|
%1 = trunc <4 x i16> %0 to <4 x i8>
|
|
store <4 x i8> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test4i8w(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test4i8w:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: addis r4, r2, .LCPI2_0@toc@ha
|
|
; CHECK-NEXT: addi r4, r4, .LCPI2_0@toc@l
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v3, vs0
|
|
; CHECK-NEXT: vperm v2, v2, v2, v3
|
|
; CHECK-NEXT: xxsldwi vs0, v2, v2, 2
|
|
; CHECK-NEXT: stfiwx f0, 0, r3
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test4i8w:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: addis r5, r2, .LCPI2_0@toc@ha
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r4, r5, .LCPI2_0@toc@l
|
|
; CHECK-BE-NEXT: lxvw4x v3, 0, r4
|
|
; CHECK-BE-NEXT: addi r4, r1, -16
|
|
; CHECK-BE-NEXT: vperm v2, v2, v2, v3
|
|
; CHECK-BE-NEXT: stxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: lwz r4, -16(r1)
|
|
; CHECK-BE-NEXT: stw r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <4 x i32>, ptr %SrcPtr, align 16
|
|
%1 = trunc <4 x i32> %0 to <4 x i8>
|
|
store <4 x i8> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test2i8(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test2i8:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: vpkuhum v2, v2, v2
|
|
; CHECK-NEXT: xxswapd vs0, v2
|
|
; CHECK-NEXT: mffprd r4, f0
|
|
; CHECK-NEXT: clrldi r4, r4, 48
|
|
; CHECK-NEXT: sth r4, 0(r3)
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test2i8:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r5, r1, -16
|
|
; CHECK-BE-NEXT: vpkuhum v2, v2, v2
|
|
; CHECK-BE-NEXT: stxvw4x v2, 0, r5
|
|
; CHECK-BE-NEXT: lhz r4, -16(r1)
|
|
; CHECK-BE-NEXT: sth r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <2 x i16>, ptr %SrcPtr, align 16
|
|
%1 = trunc <2 x i16> %0 to <2 x i8>
|
|
store <2 x i8> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test4i16(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test4i16:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: vpkuwum v2, v2, v2
|
|
; CHECK-NEXT: xxswapd vs0, v2
|
|
; CHECK-NEXT: stfdx f0, 0, r3
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test4i16:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r5, r1, -16
|
|
; CHECK-BE-NEXT: vpkuwum v2, v2, v2
|
|
; CHECK-BE-NEXT: stxvd2x v2, 0, r5
|
|
; CHECK-BE-NEXT: ld r4, -16(r1)
|
|
; CHECK-BE-NEXT: std r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <4 x i32>, ptr %SrcPtr, align 16
|
|
%1 = trunc <4 x i32> %0 to <4 x i16>
|
|
store <4 x i16> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test2i16(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test2i16:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: vpkuwum v2, v2, v2
|
|
; CHECK-NEXT: xxsldwi vs0, v2, v2, 2
|
|
; CHECK-NEXT: stfiwx f0, 0, r3
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test2i16:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r5, r1, -16
|
|
; CHECK-BE-NEXT: vpkuwum v2, v2, v2
|
|
; CHECK-BE-NEXT: stxvw4x v2, 0, r5
|
|
; CHECK-BE-NEXT: lwz r4, -16(r1)
|
|
; CHECK-BE-NEXT: stw r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <2 x i32>, ptr %SrcPtr, align 16
|
|
%1 = trunc <2 x i32> %0 to <2 x i16>
|
|
store <2 x i16> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test2i16d(ptr nocapture %Sink, ptr nocapture readonly %SrcPtr) {
|
|
; CHECK-LABEL: test2i16d:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: addis r4, r2, .LCPI6_0@toc@ha
|
|
; CHECK-NEXT: addi r4, r4, .LCPI6_0@toc@l
|
|
; CHECK-NEXT: xxswapd v2, vs0
|
|
; CHECK-NEXT: lxvd2x vs0, 0, r4
|
|
; CHECK-NEXT: xxswapd v3, vs0
|
|
; CHECK-NEXT: vperm v2, v2, v2, v3
|
|
; CHECK-NEXT: xxsldwi vs0, v2, v2, 2
|
|
; CHECK-NEXT: stfiwx f0, 0, r3
|
|
; CHECK-NEXT: blr
|
|
;
|
|
; CHECK-BE-LABEL: test2i16d:
|
|
; CHECK-BE: # %bb.0: # %entry
|
|
; CHECK-BE-NEXT: addis r5, r2, .LCPI6_0@toc@ha
|
|
; CHECK-BE-NEXT: lxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: addi r4, r5, .LCPI6_0@toc@l
|
|
; CHECK-BE-NEXT: lxvw4x v3, 0, r4
|
|
; CHECK-BE-NEXT: addi r4, r1, -16
|
|
; CHECK-BE-NEXT: vperm v2, v2, v2, v3
|
|
; CHECK-BE-NEXT: stxvw4x v2, 0, r4
|
|
; CHECK-BE-NEXT: lwz r4, -16(r1)
|
|
; CHECK-BE-NEXT: stw r4, 0(r3)
|
|
; CHECK-BE-NEXT: blr
|
|
entry:
|
|
%0 = load <2 x i64>, ptr %SrcPtr, align 16
|
|
%1 = trunc <2 x i64> %0 to <2 x i16>
|
|
store <2 x i16> %1, ptr %Sink, align 16
|
|
ret void
|
|
}
|