Files
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

346 lines
12 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
; RUN: llc -mtriple=aarch64 -verify-machineinstrs %s -o - | FileCheck %s --check-prefixes=CHECK,CHECK-SD,CHECK-SD-NOFP16
; RUN: llc -mtriple=aarch64 -mattr=+fullfp16 -verify-machineinstrs %s -o - | FileCheck %s --check-prefixes=CHECK,CHECK-SD,CHECK-SD-FP16
; RUN: llc -mtriple=aarch64 -global-isel -verify-machineinstrs %s -o - | FileCheck %s --check-prefixes=CHECK,CHECK-GI,CHECK-GI-NOFP16
; RUN: llc -mtriple=aarch64 -mattr=+fullfp16 -global-isel -verify-machineinstrs %s -o - | FileCheck %s --check-prefixes=CHECK,CHECK-GI,CHECK-GI-FP16
define double @fdiv_f64(double %a, double %b) {
; CHECK-LABEL: fdiv_f64:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: fdiv d0, d0, d1
; CHECK-NEXT: ret
entry:
%c = fdiv double %a, %b
ret double %c
}
define float @fdiv_f32(float %a, float %b) {
; CHECK-LABEL: fdiv_f32:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: fdiv s0, s0, s1
; CHECK-NEXT: ret
entry:
%c = fdiv float %a, %b
ret float %c
}
define half @fdiv_f16(half %a, half %b) {
; CHECK-SD-NOFP16-LABEL: fdiv_f16:
; CHECK-SD-NOFP16: // %bb.0: // %entry
; CHECK-SD-NOFP16-NEXT: fcvt s1, h1
; CHECK-SD-NOFP16-NEXT: fcvt s0, h0
; CHECK-SD-NOFP16-NEXT: fdiv s0, s0, s1
; CHECK-SD-NOFP16-NEXT: fcvt h0, s0
; CHECK-SD-NOFP16-NEXT: ret
;
; CHECK-SD-FP16-LABEL: fdiv_f16:
; CHECK-SD-FP16: // %bb.0: // %entry
; CHECK-SD-FP16-NEXT: fdiv h0, h0, h1
; CHECK-SD-FP16-NEXT: ret
;
; CHECK-GI-NOFP16-LABEL: fdiv_f16:
; CHECK-GI-NOFP16: // %bb.0: // %entry
; CHECK-GI-NOFP16-NEXT: fcvt s0, h0
; CHECK-GI-NOFP16-NEXT: fcvt s1, h1
; CHECK-GI-NOFP16-NEXT: fdiv s0, s0, s1
; CHECK-GI-NOFP16-NEXT: fcvt h0, s0
; CHECK-GI-NOFP16-NEXT: ret
;
; CHECK-GI-FP16-LABEL: fdiv_f16:
; CHECK-GI-FP16: // %bb.0: // %entry
; CHECK-GI-FP16-NEXT: fdiv h0, h0, h1
; CHECK-GI-FP16-NEXT: ret
entry:
%c = fdiv half %a, %b
ret half %c
}
define <2 x double> @fdiv_v2f64(<2 x double> %a, <2 x double> %b) {
; CHECK-LABEL: fdiv_v2f64:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: fdiv v0.2d, v0.2d, v1.2d
; CHECK-NEXT: ret
entry:
%c = fdiv <2 x double> %a, %b
ret <2 x double> %c
}
define <3 x double> @fdiv_v3f64(<3 x double> %a, <3 x double> %b) {
; CHECK-SD-LABEL: fdiv_v3f64:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: // kill: def $d3 killed $d3 def $q3
; CHECK-SD-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-SD-NEXT: // kill: def $d4 killed $d4 def $q4
; CHECK-SD-NEXT: // kill: def $d1 killed $d1 def $q1
; CHECK-SD-NEXT: // kill: def $d2 killed $d2 def $q2
; CHECK-SD-NEXT: // kill: def $d5 killed $d5 def $q5
; CHECK-SD-NEXT: mov v3.d[1], v4.d[0]
; CHECK-SD-NEXT: mov v0.d[1], v1.d[0]
; CHECK-SD-NEXT: fdiv v2.2d, v2.2d, v5.2d
; CHECK-SD-NEXT: // kill: def $d2 killed $d2 killed $q2
; CHECK-SD-NEXT: fdiv v0.2d, v0.2d, v3.2d
; CHECK-SD-NEXT: mov d1, v0.d[1]
; CHECK-SD-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: fdiv_v3f64:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-GI-NEXT: // kill: def $d3 killed $d3 def $q3
; CHECK-GI-NEXT: // kill: def $d1 killed $d1 def $q1
; CHECK-GI-NEXT: // kill: def $d4 killed $d4 def $q4
; CHECK-GI-NEXT: fdiv d2, d2, d5
; CHECK-GI-NEXT: mov v0.d[1], v1.d[0]
; CHECK-GI-NEXT: mov v3.d[1], v4.d[0]
; CHECK-GI-NEXT: fdiv v0.2d, v0.2d, v3.2d
; CHECK-GI-NEXT: mov d1, v0.d[1]
; CHECK-GI-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-GI-NEXT: ret
entry:
%c = fdiv <3 x double> %a, %b
ret <3 x double> %c
}
define <4 x double> @fdiv_v4f64(<4 x double> %a, <4 x double> %b) {
; CHECK-SD-LABEL: fdiv_v4f64:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: fdiv v1.2d, v1.2d, v3.2d
; CHECK-SD-NEXT: fdiv v0.2d, v0.2d, v2.2d
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: fdiv_v4f64:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: fdiv v0.2d, v0.2d, v2.2d
; CHECK-GI-NEXT: fdiv v1.2d, v1.2d, v3.2d
; CHECK-GI-NEXT: ret
entry:
%c = fdiv <4 x double> %a, %b
ret <4 x double> %c
}
define <2 x float> @fdiv_v2f32(<2 x float> %a, <2 x float> %b) {
; CHECK-LABEL: fdiv_v2f32:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: fdiv v0.2s, v0.2s, v1.2s
; CHECK-NEXT: ret
entry:
%c = fdiv <2 x float> %a, %b
ret <2 x float> %c
}
define <3 x float> @fdiv_v3f32(<3 x float> %a, <3 x float> %b) {
; CHECK-LABEL: fdiv_v3f32:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: fdiv v0.4s, v0.4s, v1.4s
; CHECK-NEXT: ret
entry:
%c = fdiv <3 x float> %a, %b
ret <3 x float> %c
}
define <4 x float> @fdiv_v4f32(<4 x float> %a, <4 x float> %b) {
; CHECK-LABEL: fdiv_v4f32:
; CHECK: // %bb.0: // %entry
; CHECK-NEXT: fdiv v0.4s, v0.4s, v1.4s
; CHECK-NEXT: ret
entry:
%c = fdiv <4 x float> %a, %b
ret <4 x float> %c
}
define <8 x float> @fdiv_v8f32(<8 x float> %a, <8 x float> %b) {
; CHECK-SD-LABEL: fdiv_v8f32:
; CHECK-SD: // %bb.0: // %entry
; CHECK-SD-NEXT: fdiv v1.4s, v1.4s, v3.4s
; CHECK-SD-NEXT: fdiv v0.4s, v0.4s, v2.4s
; CHECK-SD-NEXT: ret
;
; CHECK-GI-LABEL: fdiv_v8f32:
; CHECK-GI: // %bb.0: // %entry
; CHECK-GI-NEXT: fdiv v0.4s, v0.4s, v2.4s
; CHECK-GI-NEXT: fdiv v1.4s, v1.4s, v3.4s
; CHECK-GI-NEXT: ret
entry:
%c = fdiv <8 x float> %a, %b
ret <8 x float> %c
}
define <7 x half> @fdiv_v7f16(<7 x half> %a, <7 x half> %b) {
; CHECK-SD-NOFP16-LABEL: fdiv_v7f16:
; CHECK-SD-NOFP16: // %bb.0: // %entry
; CHECK-SD-NOFP16-NEXT: fcvtl v2.4s, v1.4h
; CHECK-SD-NOFP16-NEXT: fcvtl v3.4s, v0.4h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v1.4s, v1.8h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-SD-NOFP16-NEXT: fdiv v2.4s, v3.4s, v2.4s
; CHECK-SD-NOFP16-NEXT: fdiv v1.4s, v0.4s, v1.4s
; CHECK-SD-NOFP16-NEXT: fcvtn v0.4h, v2.4s
; CHECK-SD-NOFP16-NEXT: fcvtn2 v0.8h, v1.4s
; CHECK-SD-NOFP16-NEXT: ret
;
; CHECK-SD-FP16-LABEL: fdiv_v7f16:
; CHECK-SD-FP16: // %bb.0: // %entry
; CHECK-SD-FP16-NEXT: fdiv v0.8h, v0.8h, v1.8h
; CHECK-SD-FP16-NEXT: ret
;
; CHECK-GI-NOFP16-LABEL: fdiv_v7f16:
; CHECK-GI-NOFP16: // %bb.0: // %entry
; CHECK-GI-NOFP16-NEXT: fcvtl v2.4s, v0.4h
; CHECK-GI-NOFP16-NEXT: fcvtl v3.4s, v1.4h
; CHECK-GI-NOFP16-NEXT: mov v4.h[0], v0.h[4]
; CHECK-GI-NOFP16-NEXT: fdiv v2.4s, v2.4s, v3.4s
; CHECK-GI-NOFP16-NEXT: mov v3.h[0], v1.h[4]
; CHECK-GI-NOFP16-NEXT: mov v4.h[1], v0.h[5]
; CHECK-GI-NOFP16-NEXT: mov v3.h[1], v1.h[5]
; CHECK-GI-NOFP16-NEXT: mov v4.h[2], v0.h[6]
; CHECK-GI-NOFP16-NEXT: mov v3.h[2], v1.h[6]
; CHECK-GI-NOFP16-NEXT: fcvtl v0.4s, v4.4h
; CHECK-GI-NOFP16-NEXT: fcvtl v1.4s, v3.4h
; CHECK-GI-NOFP16-NEXT: fcvtn v2.4h, v2.4s
; CHECK-GI-NOFP16-NEXT: fdiv v1.4s, v0.4s, v1.4s
; CHECK-GI-NOFP16-NEXT: mov v0.h[0], v2.h[0]
; CHECK-GI-NOFP16-NEXT: mov v0.h[1], v2.h[1]
; CHECK-GI-NOFP16-NEXT: mov v0.h[2], v2.h[2]
; CHECK-GI-NOFP16-NEXT: fcvtn v1.4h, v1.4s
; CHECK-GI-NOFP16-NEXT: mov v0.h[3], v2.h[3]
; CHECK-GI-NOFP16-NEXT: mov v0.h[4], v1.h[0]
; CHECK-GI-NOFP16-NEXT: mov v0.h[5], v1.h[1]
; CHECK-GI-NOFP16-NEXT: mov v0.h[6], v1.h[2]
; CHECK-GI-NOFP16-NEXT: ret
;
; CHECK-GI-FP16-LABEL: fdiv_v7f16:
; CHECK-GI-FP16: // %bb.0: // %entry
; CHECK-GI-FP16-NEXT: fdiv v0.8h, v0.8h, v1.8h
; CHECK-GI-FP16-NEXT: ret
entry:
%c = fdiv <7 x half> %a, %b
ret <7 x half> %c
}
define <4 x half> @fdiv_v4f16(<4 x half> %a, <4 x half> %b) {
; CHECK-SD-NOFP16-LABEL: fdiv_v4f16:
; CHECK-SD-NOFP16: // %bb.0: // %entry
; CHECK-SD-NOFP16-NEXT: fcvtl v1.4s, v1.4h
; CHECK-SD-NOFP16-NEXT: fcvtl v0.4s, v0.4h
; CHECK-SD-NOFP16-NEXT: fdiv v0.4s, v0.4s, v1.4s
; CHECK-SD-NOFP16-NEXT: fcvtn v0.4h, v0.4s
; CHECK-SD-NOFP16-NEXT: ret
;
; CHECK-SD-FP16-LABEL: fdiv_v4f16:
; CHECK-SD-FP16: // %bb.0: // %entry
; CHECK-SD-FP16-NEXT: fdiv v0.4h, v0.4h, v1.4h
; CHECK-SD-FP16-NEXT: ret
;
; CHECK-GI-NOFP16-LABEL: fdiv_v4f16:
; CHECK-GI-NOFP16: // %bb.0: // %entry
; CHECK-GI-NOFP16-NEXT: fcvtl v0.4s, v0.4h
; CHECK-GI-NOFP16-NEXT: fcvtl v1.4s, v1.4h
; CHECK-GI-NOFP16-NEXT: fdiv v0.4s, v0.4s, v1.4s
; CHECK-GI-NOFP16-NEXT: fcvtn v0.4h, v0.4s
; CHECK-GI-NOFP16-NEXT: ret
;
; CHECK-GI-FP16-LABEL: fdiv_v4f16:
; CHECK-GI-FP16: // %bb.0: // %entry
; CHECK-GI-FP16-NEXT: fdiv v0.4h, v0.4h, v1.4h
; CHECK-GI-FP16-NEXT: ret
entry:
%c = fdiv <4 x half> %a, %b
ret <4 x half> %c
}
define <8 x half> @fdiv_v8f16(<8 x half> %a, <8 x half> %b) {
; CHECK-SD-NOFP16-LABEL: fdiv_v8f16:
; CHECK-SD-NOFP16: // %bb.0: // %entry
; CHECK-SD-NOFP16-NEXT: fcvtl v2.4s, v1.4h
; CHECK-SD-NOFP16-NEXT: fcvtl v3.4s, v0.4h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v1.4s, v1.8h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-SD-NOFP16-NEXT: fdiv v2.4s, v3.4s, v2.4s
; CHECK-SD-NOFP16-NEXT: fdiv v1.4s, v0.4s, v1.4s
; CHECK-SD-NOFP16-NEXT: fcvtn v0.4h, v2.4s
; CHECK-SD-NOFP16-NEXT: fcvtn2 v0.8h, v1.4s
; CHECK-SD-NOFP16-NEXT: ret
;
; CHECK-SD-FP16-LABEL: fdiv_v8f16:
; CHECK-SD-FP16: // %bb.0: // %entry
; CHECK-SD-FP16-NEXT: fdiv v0.8h, v0.8h, v1.8h
; CHECK-SD-FP16-NEXT: ret
;
; CHECK-GI-NOFP16-LABEL: fdiv_v8f16:
; CHECK-GI-NOFP16: // %bb.0: // %entry
; CHECK-GI-NOFP16-NEXT: fcvtl v2.4s, v0.4h
; CHECK-GI-NOFP16-NEXT: fcvtl v3.4s, v1.4h
; CHECK-GI-NOFP16-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-GI-NOFP16-NEXT: fcvtl2 v1.4s, v1.8h
; CHECK-GI-NOFP16-NEXT: fdiv v2.4s, v2.4s, v3.4s
; CHECK-GI-NOFP16-NEXT: fdiv v1.4s, v0.4s, v1.4s
; CHECK-GI-NOFP16-NEXT: fcvtn v0.4h, v2.4s
; CHECK-GI-NOFP16-NEXT: fcvtn2 v0.8h, v1.4s
; CHECK-GI-NOFP16-NEXT: ret
;
; CHECK-GI-FP16-LABEL: fdiv_v8f16:
; CHECK-GI-FP16: // %bb.0: // %entry
; CHECK-GI-FP16-NEXT: fdiv v0.8h, v0.8h, v1.8h
; CHECK-GI-FP16-NEXT: ret
entry:
%c = fdiv <8 x half> %a, %b
ret <8 x half> %c
}
define <16 x half> @fdiv_v16f16(<16 x half> %a, <16 x half> %b) {
; CHECK-SD-NOFP16-LABEL: fdiv_v16f16:
; CHECK-SD-NOFP16: // %bb.0: // %entry
; CHECK-SD-NOFP16-NEXT: fcvtl2 v4.4s, v2.8h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v5.4s, v0.8h
; CHECK-SD-NOFP16-NEXT: fcvtl v2.4s, v2.4h
; CHECK-SD-NOFP16-NEXT: fcvtl v0.4s, v0.4h
; CHECK-SD-NOFP16-NEXT: fdiv v4.4s, v5.4s, v4.4s
; CHECK-SD-NOFP16-NEXT: fcvtl v5.4s, v1.4h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v1.4s, v1.8h
; CHECK-SD-NOFP16-NEXT: fdiv v0.4s, v0.4s, v2.4s
; CHECK-SD-NOFP16-NEXT: fcvtl v2.4s, v3.4h
; CHECK-SD-NOFP16-NEXT: fcvtl2 v3.4s, v3.8h
; CHECK-SD-NOFP16-NEXT: fdiv v2.4s, v5.4s, v2.4s
; CHECK-SD-NOFP16-NEXT: fcvtn v0.4h, v0.4s
; CHECK-SD-NOFP16-NEXT: fcvtn2 v0.8h, v4.4s
; CHECK-SD-NOFP16-NEXT: fdiv v3.4s, v1.4s, v3.4s
; CHECK-SD-NOFP16-NEXT: fcvtn v1.4h, v2.4s
; CHECK-SD-NOFP16-NEXT: fcvtn2 v1.8h, v3.4s
; CHECK-SD-NOFP16-NEXT: ret
;
; CHECK-SD-FP16-LABEL: fdiv_v16f16:
; CHECK-SD-FP16: // %bb.0: // %entry
; CHECK-SD-FP16-NEXT: fdiv v1.8h, v1.8h, v3.8h
; CHECK-SD-FP16-NEXT: fdiv v0.8h, v0.8h, v2.8h
; CHECK-SD-FP16-NEXT: ret
;
; CHECK-GI-NOFP16-LABEL: fdiv_v16f16:
; CHECK-GI-NOFP16: // %bb.0: // %entry
; CHECK-GI-NOFP16-NEXT: fcvtl v4.4s, v0.4h
; CHECK-GI-NOFP16-NEXT: fcvtl v5.4s, v2.4h
; CHECK-GI-NOFP16-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-GI-NOFP16-NEXT: fcvtl2 v2.4s, v2.8h
; CHECK-GI-NOFP16-NEXT: fdiv v4.4s, v4.4s, v5.4s
; CHECK-GI-NOFP16-NEXT: fcvtl v5.4s, v3.4h
; CHECK-GI-NOFP16-NEXT: fdiv v2.4s, v0.4s, v2.4s
; CHECK-GI-NOFP16-NEXT: fcvtl v0.4s, v1.4h
; CHECK-GI-NOFP16-NEXT: fdiv v5.4s, v0.4s, v5.4s
; CHECK-GI-NOFP16-NEXT: fcvtl2 v0.4s, v1.8h
; CHECK-GI-NOFP16-NEXT: fcvtl2 v1.4s, v3.8h
; CHECK-GI-NOFP16-NEXT: fdiv v3.4s, v0.4s, v1.4s
; CHECK-GI-NOFP16-NEXT: fcvtn v0.4h, v4.4s
; CHECK-GI-NOFP16-NEXT: fcvtn v1.4h, v5.4s
; CHECK-GI-NOFP16-NEXT: fcvtn2 v0.8h, v2.4s
; CHECK-GI-NOFP16-NEXT: fcvtn2 v1.8h, v3.4s
; CHECK-GI-NOFP16-NEXT: ret
;
; CHECK-GI-FP16-LABEL: fdiv_v16f16:
; CHECK-GI-FP16: // %bb.0: // %entry
; CHECK-GI-FP16-NEXT: fdiv v0.8h, v0.8h, v2.8h
; CHECK-GI-FP16-NEXT: fdiv v1.8h, v1.8h, v3.8h
; CHECK-GI-FP16-NEXT: ret
entry:
%c = fdiv <16 x half> %a, %b
ret <16 x half> %c
}