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
108 lines
4.4 KiB
LLVM
108 lines
4.4 KiB
LLVM
; RUN: llc -amdgpu-scalarize-global-loads=false -mtriple=amdgcn -mcpu=tahiti < %s | FileCheck -check-prefix=SI %s
|
|
; RUN: llc -amdgpu-scalarize-global-loads=false -mtriple=amdgcn -mcpu=tonga < %s | FileCheck -check-prefix=SI %s
|
|
|
|
declare double @llvm.fabs.f64(double) #0
|
|
|
|
; SI-LABEL: {{^}}fsub_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @fsub_f64(ptr addrspace(1) %out, ptr addrspace(1) %in1,
|
|
ptr addrspace(1) %in2) {
|
|
%r0 = load double, ptr addrspace(1) %in1
|
|
%r1 = load double, ptr addrspace(1) %in2
|
|
%r2 = fsub double %r0, %r1
|
|
store double %r2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}fsub_fabs_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|}}
|
|
define amdgpu_kernel void @fsub_fabs_f64(ptr addrspace(1) %out, ptr addrspace(1) %in1,
|
|
ptr addrspace(1) %in2) {
|
|
%r0 = load double, ptr addrspace(1) %in1
|
|
%r1 = load double, ptr addrspace(1) %in2
|
|
%r1.fabs = call double @llvm.fabs.f64(double %r1) #0
|
|
%r2 = fsub double %r0, %r1.fabs
|
|
store double %r2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}fsub_fabs_inv_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], |v\[[0-9]+:[0-9]+\]|, -v\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @fsub_fabs_inv_f64(ptr addrspace(1) %out, ptr addrspace(1) %in1,
|
|
ptr addrspace(1) %in2) {
|
|
%r0 = load double, ptr addrspace(1) %in1
|
|
%r1 = load double, ptr addrspace(1) %in2
|
|
%r0.fabs = call double @llvm.fabs.f64(double %r0) #0
|
|
%r2 = fsub double %r0.fabs, %r1
|
|
store double %r2, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}s_fsub_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @s_fsub_f64(ptr addrspace(1) %out, double %a, double %b) {
|
|
%sub = fsub double %a, %b
|
|
store double %sub, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}s_fsub_imm_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], -s\[[0-9]+:[0-9]+\]}}, 4.0
|
|
define amdgpu_kernel void @s_fsub_imm_f64(ptr addrspace(1) %out, double %a, double %b) {
|
|
%sub = fsub double 4.0, %a
|
|
store double %sub, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}s_fsub_imm_inv_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\]}}, -4.0
|
|
define amdgpu_kernel void @s_fsub_imm_inv_f64(ptr addrspace(1) %out, double %a, double %b) {
|
|
%sub = fsub double %a, 4.0
|
|
store double %sub, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}s_fsub_self_f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -s\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @s_fsub_self_f64(ptr addrspace(1) %out, double %a) {
|
|
%sub = fsub double %a, %a
|
|
store double %sub, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}fsub_v2f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @fsub_v2f64(ptr addrspace(1) %out, <2 x double> %a, <2 x double> %b) {
|
|
%sub = fsub <2 x double> %a, %b
|
|
store <2 x double> %sub, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}fsub_v4f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @fsub_v4f64(ptr addrspace(1) %out, ptr addrspace(1) %in) {
|
|
%b_ptr = getelementptr <4 x double>, ptr addrspace(1) %in, i32 1
|
|
%a = load <4 x double>, ptr addrspace(1) %in
|
|
%b = load <4 x double>, ptr addrspace(1) %b_ptr
|
|
%result = fsub <4 x double> %a, %b
|
|
store <4 x double> %result, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; SI-LABEL: {{^}}s_fsub_v4f64:
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
; SI: v_add_f64 {{v\[[0-9]+:[0-9]+\], s\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
|
|
define amdgpu_kernel void @s_fsub_v4f64(ptr addrspace(1) %out, <4 x double> %a, <4 x double> %b) {
|
|
%result = fsub <4 x double> %a, %b
|
|
store <4 x double> %result, ptr addrspace(1) %out, align 16
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { nounwind readnone }
|