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
114 lines
3.1 KiB
LLVM
114 lines
3.1 KiB
LLVM
; RUN: llc -mtriple=arm -mattr=+neon %s -o - | FileCheck %s
|
|
|
|
; This tests icmp operations that do not map directly to NEON instructions.
|
|
; Not-equal (ne) operations are implemented by VCEQ/VMVN. Less-than (lt/ult)
|
|
; and less-than-or-equal (le/ule) are implemented by swapping the arguments
|
|
; to VCGT and VCGE. Test all the operand types for not-equal but only sample
|
|
; the other operations.
|
|
|
|
define <8 x i8> @vcnei8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcnei8:
|
|
;CHECK: vceq.i8
|
|
;CHECK-NEXT: vmvn
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = icmp ne <8 x i8> %tmp1, %tmp2
|
|
%tmp4 = sext <8 x i1> %tmp3 to <8 x i8>
|
|
ret <8 x i8> %tmp4
|
|
}
|
|
|
|
define <4 x i16> @vcnei16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcnei16:
|
|
;CHECK: vceq.i16
|
|
;CHECK-NEXT: vmvn
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = icmp ne <4 x i16> %tmp1, %tmp2
|
|
%tmp4 = sext <4 x i1> %tmp3 to <4 x i16>
|
|
ret <4 x i16> %tmp4
|
|
}
|
|
|
|
define <2 x i32> @vcnei32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcnei32:
|
|
;CHECK: vceq.i32
|
|
;CHECK-NEXT: vmvn
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = icmp ne <2 x i32> %tmp1, %tmp2
|
|
%tmp4 = sext <2 x i1> %tmp3 to <2 x i32>
|
|
ret <2 x i32> %tmp4
|
|
}
|
|
|
|
define <16 x i8> @vcneQi8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcneQi8:
|
|
;CHECK: vceq.i8
|
|
;CHECK-NEXT: vmvn
|
|
%tmp1 = load <16 x i8>, ptr %A
|
|
%tmp2 = load <16 x i8>, ptr %B
|
|
%tmp3 = icmp ne <16 x i8> %tmp1, %tmp2
|
|
%tmp4 = sext <16 x i1> %tmp3 to <16 x i8>
|
|
ret <16 x i8> %tmp4
|
|
}
|
|
|
|
define <8 x i16> @vcneQi16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcneQi16:
|
|
;CHECK: vceq.i16
|
|
;CHECK-NEXT: vmvn
|
|
%tmp1 = load <8 x i16>, ptr %A
|
|
%tmp2 = load <8 x i16>, ptr %B
|
|
%tmp3 = icmp ne <8 x i16> %tmp1, %tmp2
|
|
%tmp4 = sext <8 x i1> %tmp3 to <8 x i16>
|
|
ret <8 x i16> %tmp4
|
|
}
|
|
|
|
define <4 x i32> @vcneQi32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcneQi32:
|
|
;CHECK: vceq.i32
|
|
;CHECK-NEXT: vmvn
|
|
%tmp1 = load <4 x i32>, ptr %A
|
|
%tmp2 = load <4 x i32>, ptr %B
|
|
%tmp3 = icmp ne <4 x i32> %tmp1, %tmp2
|
|
%tmp4 = sext <4 x i1> %tmp3 to <4 x i32>
|
|
ret <4 x i32> %tmp4
|
|
}
|
|
|
|
define <16 x i8> @vcltQs8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcltQs8:
|
|
;CHECK: vcgt.s8
|
|
%tmp1 = load <16 x i8>, ptr %A
|
|
%tmp2 = load <16 x i8>, ptr %B
|
|
%tmp3 = icmp slt <16 x i8> %tmp1, %tmp2
|
|
%tmp4 = sext <16 x i1> %tmp3 to <16 x i8>
|
|
ret <16 x i8> %tmp4
|
|
}
|
|
|
|
define <4 x i16> @vcles16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcles16:
|
|
;CHECK: vcge.s16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = icmp sle <4 x i16> %tmp1, %tmp2
|
|
%tmp4 = sext <4 x i1> %tmp3 to <4 x i16>
|
|
ret <4 x i16> %tmp4
|
|
}
|
|
|
|
define <4 x i16> @vcltu16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcltu16:
|
|
;CHECK: vcgt.u16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = icmp ult <4 x i16> %tmp1, %tmp2
|
|
%tmp4 = sext <4 x i1> %tmp3 to <4 x i16>
|
|
ret <4 x i16> %tmp4
|
|
}
|
|
|
|
define <4 x i32> @vcleQu32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vcleQu32:
|
|
;CHECK: vcge.u32
|
|
%tmp1 = load <4 x i32>, ptr %A
|
|
%tmp2 = load <4 x i32>, ptr %B
|
|
%tmp3 = icmp ule <4 x i32> %tmp1, %tmp2
|
|
%tmp4 = sext <4 x i1> %tmp3 to <4 x i32>
|
|
ret <4 x i32> %tmp4
|
|
}
|