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
39 lines
2.0 KiB
LLVM
39 lines
2.0 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: llc < %s -mcpu=sm_100 -mattr=+ptx88 -O3 -disable-post-ra \
|
|
; RUN: -frame-pointer=all -verify-machineinstrs \
|
|
; RUN: | FileCheck %s --check-prefixes=CHECK-F32X2
|
|
; RUN: %if ptxas-sm_100 && ptxas-isa-8.8 %{ \
|
|
; RUN: llc < %s -mcpu=sm_100 -mattr=+ptx88 -O3 -disable-post-ra \
|
|
; RUN: -frame-pointer=all -verify-machineinstrs | %ptxas-verify -arch=sm_100 \
|
|
; RUN: %}
|
|
target triple = "nvptx64-nvidia-cuda"
|
|
|
|
; Since fdiv doesn't support f32x2, this will create BUILD_VECTORs that will be
|
|
; folded into the store, turning it into st.global.v8.b32.
|
|
define void @writevec(<8 x float> %v1, <8 x float> %v2, ptr addrspace(1) %p) {
|
|
; CHECK-F32X2-LABEL: writevec(
|
|
; CHECK-F32X2: {
|
|
; CHECK-F32X2-NEXT: .reg .b32 %r<25>;
|
|
; CHECK-F32X2-NEXT: .reg .b64 %rd<2>;
|
|
; CHECK-F32X2-EMPTY:
|
|
; CHECK-F32X2-NEXT: // %bb.0:
|
|
; CHECK-F32X2-NEXT: ld.param.v4.b32 {%r1, %r2, %r3, %r4}, [writevec_param_0];
|
|
; CHECK-F32X2-NEXT: ld.param.v4.b32 {%r5, %r6, %r7, %r8}, [writevec_param_0+16];
|
|
; CHECK-F32X2-NEXT: ld.param.v4.b32 {%r9, %r10, %r11, %r12}, [writevec_param_1+16];
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r13, %r8, %r12;
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r14, %r7, %r11;
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r15, %r6, %r10;
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r16, %r5, %r9;
|
|
; CHECK-F32X2-NEXT: ld.param.v4.b32 {%r17, %r18, %r19, %r20}, [writevec_param_1];
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r21, %r4, %r20;
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r22, %r3, %r19;
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r23, %r2, %r18;
|
|
; CHECK-F32X2-NEXT: div.rn.f32 %r24, %r1, %r17;
|
|
; CHECK-F32X2-NEXT: ld.param.b64 %rd1, [writevec_param_2];
|
|
; CHECK-F32X2-NEXT: st.global.v8.b32 [%rd1], {%r24, %r23, %r22, %r21, %r16, %r15, %r14, %r13};
|
|
; CHECK-F32X2-NEXT: ret;
|
|
%v = fdiv <8 x float> %v1, %v2
|
|
store <8 x float> %v, ptr addrspace(1) %p, align 32
|
|
ret void
|
|
}
|