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
328 lines
16 KiB
LLVM
328 lines
16 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 4
|
|
; RUN: opt -passes="print<cost-model>" -cost-kind=all 2>&1 -disable-output -mtriple=arm64-apple-macosx < %s | FileCheck %s
|
|
|
|
define void @vec3_i32(<3 x i32> %a, <3 x i32> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i32'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i32>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i32> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %cmp = icmp uge <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sel = select <3 x i1> %cmp, <3 x i32> %add, <3 x i32> %sub
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i32> %sel, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i32>, ptr %src, align 1
|
|
%add = add <3 x i32> %l, %b
|
|
%cmp = icmp uge <3 x i32> %add, %a
|
|
%sub = sub <3 x i32> %add, %a
|
|
%sel = select <3 x i1> %cmp, <3 x i32> %add, <3 x i32> %sub
|
|
store <3 x i32> %sel, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i32_default_alignment(<3 x i32> %a, <3 x i32> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i32_default_alignment'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:1 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i32>, ptr %src, align 16
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i32> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %cmp = icmp uge <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sel = select <3 x i1> %cmp, <3 x i32> %add, <3 x i32> %sub
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: store <3 x i32> %sel, ptr %dst, align 16
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i32>, ptr %src
|
|
%add = add <3 x i32> %l, %b
|
|
%cmp = icmp uge <3 x i32> %add, %a
|
|
%sub = sub <3 x i32> %add, %a
|
|
%sel = select <3 x i1> %cmp, <3 x i32> %add, <3 x i32> %sub
|
|
store <3 x i32> %sel, ptr %dst
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i16(<3 x i16> %a, <3 x i16> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i16'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i16>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i16> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %cmp = icmp uge <3 x i16> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i16> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sel = select <3 x i1> %cmp, <3 x i16> %add, <3 x i16> %sub
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i16> %sel, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i16>, ptr %src, align 1
|
|
%add = add <3 x i16> %l, %b
|
|
%cmp = icmp uge <3 x i16> %add, %a
|
|
%sub = sub <3 x i16> %add, %a
|
|
%sel = select <3 x i1> %cmp, <3 x i16> %add, <3 x i16> %sub
|
|
store <3 x i16> %sel, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec7_i16(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec7_i16'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <7 x i16>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <7 x i16> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:1 SizeLat:1 for: store <7 x i16> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <7 x i16>, ptr %src, align 1
|
|
%add = add <7 x i16> %l, %l
|
|
store <7 x i16> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec6_i16(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec6_i16'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <6 x i16>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <6 x i16> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <6 x i16> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <6 x i16>, ptr %src, align 1
|
|
%add = add <6 x i16> %l, %l
|
|
store <6 x i16> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec5_i16(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec5_i16'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <5 x i16>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <5 x i16> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <5 x i16> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <5 x i16>, ptr %src, align 1
|
|
%add = add <5 x i16> %l, %l
|
|
store <5 x i16> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i16_zext_i32(<3 x i32> %a, <3 x i32> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i16_zext_i32'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i16>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %l.ext = zext <3 x i16> %l to <3 x i32>
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i32> %l.ext, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub.trunc = trunc <3 x i32> %sub to <3 x i16>
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i16> %sub.trunc, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i16>, ptr %src, align 1
|
|
%l.ext = zext <3 x i16> %l to <3 x i32>
|
|
%add = add <3 x i32> %l.ext, %b
|
|
%sub = sub <3 x i32> %add, %a
|
|
%sub.trunc = trunc <3 x i32> %sub to <3 x i16>
|
|
store <3 x i16> %sub.trunc, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i8(<3 x i8> %a, <3 x i8> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i8> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %cmp = icmp uge <3 x i8> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i8> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sel = select <3 x i1> %cmp, <3 x i8> %add, <3 x i8> %sub
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i8> %sel, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i8>, ptr %src, align 1
|
|
%add = add <3 x i8> %l, %b
|
|
%cmp = icmp uge <3 x i8> %add, %a
|
|
%sub = sub <3 x i8> %add, %a
|
|
%sel = select <3 x i1> %cmp, <3 x i8> %add, <3 x i8> %sub
|
|
store <3 x i8> %sel, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i8_zext_i32(<3 x i32> %a, <3 x i32> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i8_zext_i32'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %l.ext = zext <3 x i8> %l to <3 x i32>
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i32> %l.ext, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub.trunc = trunc <3 x i32> %sub to <3 x i8>
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i8> %sub.trunc, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i8>, ptr %src, align 1
|
|
%l.ext = zext <3 x i8> %l to <3 x i32>
|
|
%add = add <3 x i32> %l.ext, %b
|
|
%sub = sub <3 x i32> %add, %a
|
|
%sub.trunc = trunc <3 x i32> %sub to <3 x i8>
|
|
store <3 x i8> %sub.trunc, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i8_sext_i32(<3 x i32> %a, <3 x i32> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i8_sext_i32'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %l.ext = sext <3 x i8> %l to <3 x i32>
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i32> %l.ext, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i32> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub.trunc = trunc <3 x i32> %sub to <3 x i8>
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i8> %sub.trunc, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i8>, ptr %src, align 1
|
|
%l.ext = sext <3 x i8> %l to <3 x i32>
|
|
%add = add <3 x i32> %l.ext, %b
|
|
%sub = sub <3 x i32> %add, %a
|
|
%sub.trunc = trunc <3 x i32> %sub to <3 x i8>
|
|
store <3 x i8> %sub.trunc, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_i30(<3 x i30> %a, <3 x i30> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_i30'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x i30>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <3 x i30> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <3 x i30> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x i30> %sub, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x i30>, ptr %src, align 1
|
|
%add = add <3 x i30> %l, %b
|
|
%sub = sub <3 x i30> %add, %a
|
|
store <3 x i30> %sub, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_float(<3 x float> %a, <3 x float> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_float'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x float>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:1 CodeSize:1 Lat:3 SizeLat:1 for: %add = fadd <3 x float> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:1 CodeSize:1 Lat:3 SizeLat:1 for: %sub = fsub <3 x float> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x float> %sub, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x float>, ptr %src, align 1
|
|
%add = fadd <3 x float> %l, %b
|
|
%sub = fsub <3 x float> %add, %a
|
|
store <3 x float> %sub, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec3_half(<3 x half> %a, <3 x half> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec3_half'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <3 x half>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:4 CodeSize:1 Lat:3 SizeLat:1 for: %add = fadd <3 x half> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:4 CodeSize:1 Lat:3 SizeLat:1 for: %sub = fsub <3 x half> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <3 x half> %sub, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <3 x half>, ptr %src, align 1
|
|
%add = fadd <3 x half> %l, %b
|
|
%sub = fsub <3 x half> %add, %a
|
|
store <3 x half> %sub, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec15_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec15_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:4 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <15 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <15 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:4 CodeSize:1 Lat:1 SizeLat:1 for: store <15 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <15 x i8>, ptr %src, align 1
|
|
%add = add <15 x i8> %l, %l
|
|
store <15 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec14_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec14_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <14 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <14 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:1 SizeLat:1 for: store <14 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <14 x i8>, ptr %src, align 1
|
|
%add = add <14 x i8> %l, %l
|
|
store <14 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec13_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec13_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <13 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <13 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:1 SizeLat:1 for: store <13 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <13 x i8>, ptr %src, align 1
|
|
%add = add <13 x i8> %l, %l
|
|
store <13 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec12_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec12_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <12 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <12 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <12 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <12 x i8>, ptr %src, align 1
|
|
%add = add <12 x i8> %l, %l
|
|
store <12 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec11_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec11_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <11 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <11 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:1 SizeLat:1 for: store <11 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <11 x i8>, ptr %src, align 1
|
|
%add = add <11 x i8> %l, %l
|
|
store <11 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec7_i8(<7 x i8> %a, <7 x i8> %b, ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec7_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <7 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <7 x i8> %l, %b
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %sub = sub <7 x i8> %add, %a
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:3 CodeSize:1 Lat:1 SizeLat:1 for: store <7 x i8> %sub, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <7 x i8>, ptr %src, align 1
|
|
%add = add <7 x i8> %l, %b
|
|
%sub = sub <7 x i8> %add, %a
|
|
store <7 x i8> %sub, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec6_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec6_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <6 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <6 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <6 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <6 x i8>, ptr %src, align 1
|
|
%add = add <6 x i8> %l, %l
|
|
store <6 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @vec5_i8(ptr %src, ptr %dst) {
|
|
; CHECK-LABEL: 'vec5_i8'
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:4 SizeLat:1 for: %l = load <5 x i8>, ptr %src, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of 1 for: %add = add <5 x i8> %l, %l
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:2 CodeSize:1 Lat:1 SizeLat:1 for: store <5 x i8> %add, ptr %dst, align 1
|
|
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
|
|
;
|
|
%l = load <5 x i8>, ptr %src, align 1
|
|
%add = add <5 x i8> %l, %l
|
|
store <5 x i8> %add, ptr %dst, align 1
|
|
ret void
|
|
}
|
|
|
|
|