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
160 lines
5.5 KiB
LLVM
160 lines
5.5 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl,+avx512dq -O2 | FileCheck %s --check-prefix=AVX512-ALL
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f -O2| FileCheck %s --check-prefix=AVX512-ONLY
|
|
|
|
define void @load_v1i2_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i2_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i2_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i2>, ptr %a0
|
|
%d1 = trunc <1 x i2> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
define void @load_v1i3_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i3_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i3_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i3>, ptr %a0
|
|
%d1 = trunc <1 x i3> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
define void @load_v1i4_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i4_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i4_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i4>, ptr %a0
|
|
%d1 = trunc <1 x i4> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
define void @load_v1i8_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i8_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i8_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i8>, ptr %a0
|
|
%d1 = trunc <1 x i8> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
define void @load_v1i16_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i16_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i16_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i16>, ptr %a0
|
|
%d1 = trunc <1 x i16> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
define void @load_v1i32_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i32_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i32_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i32>, ptr %a0
|
|
%d1 = trunc <1 x i32> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
define void @load_v1i64_trunc_v1i1_store(ptr %a0,ptr %a1) {
|
|
; AVX512-ALL-LABEL: load_v1i64_trunc_v1i1_store:
|
|
; AVX512-ALL: # %bb.0:
|
|
; AVX512-ALL-NEXT: kmovb (%rdi), %k0
|
|
; AVX512-ALL-NEXT: kshiftlb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kshiftrb $7, %k0, %k0
|
|
; AVX512-ALL-NEXT: kmovb %k0, (%rsi)
|
|
; AVX512-ALL-NEXT: retq
|
|
;
|
|
; AVX512-ONLY-LABEL: load_v1i64_trunc_v1i1_store:
|
|
; AVX512-ONLY: # %bb.0:
|
|
; AVX512-ONLY-NEXT: movzbl (%rdi), %eax
|
|
; AVX512-ONLY-NEXT: andl $1, %eax
|
|
; AVX512-ONLY-NEXT: kmovw %eax, %k0
|
|
; AVX512-ONLY-NEXT: kmovw %k0, %eax
|
|
; AVX512-ONLY-NEXT: movb %al, (%rsi)
|
|
; AVX512-ONLY-NEXT: retq
|
|
%d0 = load <1 x i64>, ptr %a0
|
|
%d1 = trunc <1 x i64> %d0 to <1 x i1>
|
|
store <1 x i1> %d1, ptr %a1
|
|
ret void
|
|
}
|
|
|