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
173 lines
7.1 KiB
LLVM
173 lines
7.1 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
|
|
; RUN: opt -passes=instcombine %s -S | FileCheck %s
|
|
; RUN: opt -passes=instcombine -use-constant-int-for-fixed-length-splat %s -S | FileCheck %s
|
|
|
|
define ptr @vector_splat_indices_v2i64_ext0(ptr %a) {
|
|
; CHECK-LABEL: @vector_splat_indices_v2i64_ext0(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr i8, ptr [[A:%.*]], i64 16
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr i32, ptr %a, <2 x i64> <i64 4, i64 4>
|
|
%res = extractelement <2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
define ptr @vector_splat_indices_nxv2i64_ext0_inbounds(ptr %a) {
|
|
; CHECK-LABEL: @vector_splat_indices_nxv2i64_ext0_inbounds(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr inbounds nuw i8, ptr [[A:%.*]], i64 16
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%tmp = insertelement <vscale x 2 x i64> poison, i64 4, i32 0
|
|
%splatof4 = shufflevector <vscale x 2 x i64> %tmp, <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
|
|
%gep = getelementptr inbounds i32, ptr %a, <vscale x 2 x i64> %splatof4
|
|
%res = extractelement <vscale x 2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
define ptr @vector_splat_indices_nxv2i64_ext0_nuw(ptr %a) {
|
|
; CHECK-LABEL: @vector_splat_indices_nxv2i64_ext0_nuw(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr nuw i8, ptr [[A:%.*]], i64 16
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%tmp = insertelement <vscale x 2 x i64> poison, i64 4, i32 0
|
|
%splatof4 = shufflevector <vscale x 2 x i64> %tmp, <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
|
|
%gep = getelementptr nuw i32, ptr %a, <vscale x 2 x i64> %splatof4
|
|
%res = extractelement <vscale x 2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
define ptr @vector_splat_indices_nxv2i64_ext0_nusw(ptr %a) {
|
|
; CHECK-LABEL: @vector_splat_indices_nxv2i64_ext0_nusw(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr nusw nuw i8, ptr [[A:%.*]], i64 16
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%tmp = insertelement <vscale x 2 x i64> poison, i64 4, i32 0
|
|
%splatof4 = shufflevector <vscale x 2 x i64> %tmp, <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
|
|
%gep = getelementptr nusw i32, ptr %a, <vscale x 2 x i64> %splatof4
|
|
%res = extractelement <vscale x 2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
define ptr @vector_indices_v2i64_ext0(ptr %a, <2 x i64> %indices) {
|
|
; CHECK-LABEL: @vector_indices_v2i64_ext0(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = extractelement <2 x i64> [[INDICES:%.*]], i64 0
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], i64 [[TMP1]]
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr i32, ptr %a, <2 x i64> %indices
|
|
%res = extractelement <2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
define ptr @vector_indices_nxv1i64_ext0(ptr %a, <vscale x 1 x i64> %indices) {
|
|
; CHECK-LABEL: @vector_indices_nxv1i64_ext0(
|
|
; CHECK-NEXT: [[TMP1:%.*]] = extractelement <vscale x 1 x i64> [[INDICES:%.*]], i64 0
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], i64 [[TMP1]]
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr i32, ptr %a, <vscale x 1 x i64> %indices
|
|
%res = extractelement <vscale x 1 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
|
|
define ptr @vector_splat_ptrs_v2i64_ext0(ptr %a, i64 %index) {
|
|
; CHECK-LABEL: @vector_splat_ptrs_v2i64_ext0(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], i64 [[INDEX:%.*]]
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%tmp = insertelement <2 x ptr> poison, ptr %a, i32 0
|
|
%splatofa = shufflevector <2 x ptr> %tmp, <2 x ptr> poison, <2 x i32> zeroinitializer
|
|
%gep = getelementptr i32, <2 x ptr> %splatofa, i64 %index
|
|
%res = extractelement <2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
|
|
define ptr @vector_splat_ptrs_nxv2i64_ext0(ptr %a, i64 %index) {
|
|
; CHECK-LABEL: @vector_splat_ptrs_nxv2i64_ext0(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], i64 [[INDEX:%.*]]
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%tmp = insertelement <vscale x 2 x ptr> poison, ptr %a, i32 0
|
|
%splatofa = shufflevector <vscale x 2 x ptr> %tmp, <vscale x 2 x ptr> poison, <vscale x 2 x i32> zeroinitializer
|
|
%gep = getelementptr i32, <vscale x 2 x ptr> %splatofa, i64 %index
|
|
%res = extractelement <vscale x 2 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|
|
|
|
|
|
define ptr @vector_struct1_splat_indices_v4i64_ext1(ptr %a) {
|
|
; CHECK-LABEL: @vector_struct1_splat_indices_v4i64_ext1(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr i8, ptr [[A:%.*]], i64 32
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr {float, float}, ptr %a, <4 x i32> <i32 4, i32 4, i32 4, i32 4>, i32 0
|
|
%res = extractelement <4 x ptr> %gep, i32 1
|
|
ret ptr %res
|
|
}
|
|
|
|
|
|
define ptr @vector_struct2_splat_indices_v4i64_ext1(ptr %a) {
|
|
; CHECK-LABEL: @vector_struct2_splat_indices_v4i64_ext1(
|
|
; CHECK-NEXT: [[RES:%.*]] = getelementptr i8, ptr [[A:%.*]], i64 92
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr {float, [8 x float]}, ptr %a, i32 2, i32 1, <4 x i32> <i32 4, i32 4, i32 4, i32 4>
|
|
%res = extractelement <4 x ptr> %gep, i32 1
|
|
ret ptr %res
|
|
}
|
|
|
|
|
|
; Negative tests
|
|
|
|
define ptr @vector_indices_nxv2i64_ext3(ptr %a, <vscale x 2 x i64> %indices) {
|
|
; CHECK-LABEL: @vector_indices_nxv2i64_ext3(
|
|
; CHECK-NEXT: [[GEP:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], <vscale x 2 x i64> [[INDICES:%.*]]
|
|
; CHECK-NEXT: [[RES:%.*]] = extractelement <vscale x 2 x ptr> [[GEP]], i64 3
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr i32, ptr %a, <vscale x 2 x i64> %indices
|
|
%res = extractelement <vscale x 2 x ptr> %gep, i32 3
|
|
ret ptr %res
|
|
}
|
|
|
|
define ptr @vector_indices_nxv2i64_extN(ptr %a, <vscale x 2 x i64> %indices, i32 %N) {
|
|
; CHECK-LABEL: @vector_indices_nxv2i64_extN(
|
|
; CHECK-NEXT: [[GEP:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], <vscale x 2 x i64> [[INDICES:%.*]]
|
|
; CHECK-NEXT: [[RES:%.*]] = extractelement <vscale x 2 x ptr> [[GEP]], i32 [[N:%.*]]
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr i32, ptr %a, <vscale x 2 x i64> %indices
|
|
%res = extractelement <vscale x 2 x ptr> %gep, i32 %N
|
|
ret ptr %res
|
|
}
|
|
|
|
define void @vector_indices_nxv2i64_mulitple_use(ptr %a, <vscale x 2 x i64> %indices, ptr %b, ptr %c) {
|
|
; CHECK-LABEL: @vector_indices_nxv2i64_mulitple_use(
|
|
; CHECK-NEXT: [[GEP:%.*]] = getelementptr [4 x i8], ptr [[A:%.*]], <vscale x 2 x i64> [[INDICES:%.*]]
|
|
; CHECK-NEXT: [[LANE0:%.*]] = extractelement <vscale x 2 x ptr> [[GEP]], i64 0
|
|
; CHECK-NEXT: [[LANE1:%.*]] = extractelement <vscale x 2 x ptr> [[GEP]], i64 1
|
|
; CHECK-NEXT: store ptr [[LANE0]], ptr [[B:%.*]], align 8
|
|
; CHECK-NEXT: store ptr [[LANE1]], ptr [[C:%.*]], align 8
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%gep = getelementptr i32, ptr %a, <vscale x 2 x i64> %indices
|
|
%lane0 = extractelement <vscale x 2 x ptr> %gep, i32 0
|
|
%lane1 = extractelement <vscale x 2 x ptr> %gep, i32 1
|
|
store ptr %lane0, ptr %b, align 8
|
|
store ptr %lane1, ptr %c, align 8
|
|
ret void
|
|
}
|
|
|
|
define ptr @vector_ptrs_and_indices_ext0(<vscale x 4 x ptr> %a, <vscale x 4 x i64> %indices) {
|
|
; CHECK-LABEL: @vector_ptrs_and_indices_ext0(
|
|
; CHECK-NEXT: [[GEP:%.*]] = getelementptr [4 x i8], <vscale x 4 x ptr> [[A:%.*]], <vscale x 4 x i64> [[INDICES:%.*]]
|
|
; CHECK-NEXT: [[RES:%.*]] = extractelement <vscale x 4 x ptr> [[GEP]], i64 0
|
|
; CHECK-NEXT: ret ptr [[RES]]
|
|
;
|
|
%gep = getelementptr i32, <vscale x 4 x ptr> %a, <vscale x 4 x i64> %indices
|
|
%res = extractelement <vscale x 4 x ptr> %gep, i32 0
|
|
ret ptr %res
|
|
}
|