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
38 lines
1.6 KiB
LLVM
38 lines
1.6 KiB
LLVM
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_30 -mattr=+ptx60 | FileCheck %s
|
|
; RUN: %if ptxas-isa-6.0 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_30 -mattr=+ptx60 | %ptxas-verify %}
|
|
|
|
declare i32 @llvm.nvvm.fns(i32, i32, i32)
|
|
|
|
; CHECK-LABEL: .func{{.*}}fns
|
|
define i32 @fns(i32 %mask, i32 %base, i32 %offset) {
|
|
; CHECK: ld.param.b32 [[MASK:%r[0-9]+]], [fns_param_0];
|
|
; CHECK: ld.param.b32 [[BASE:%r[0-9]+]], [fns_param_1];
|
|
; CHECK: ld.param.b32 [[OFFSET:%r[0-9]+]], [fns_param_2];
|
|
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, [[MASK]], [[BASE]], [[OFFSET]];
|
|
%r0 = call i32 @llvm.nvvm.fns(i32 %mask, i32 %base, i32 %offset);
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, [[MASK]], [[BASE]], 0;
|
|
%r1 = call i32 @llvm.nvvm.fns(i32 %mask, i32 %base, i32 0);
|
|
%r01 = add i32 %r0, %r1;
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, [[MASK]], 1, [[OFFSET]];
|
|
%r2 = call i32 @llvm.nvvm.fns(i32 %mask, i32 1, i32 %offset);
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, [[MASK]], 1, 0;
|
|
%r3 = call i32 @llvm.nvvm.fns(i32 %mask, i32 1, i32 0);
|
|
%r23 = add i32 %r2, %r3;
|
|
%r0123 = add i32 %r01, %r23;
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, 2, [[BASE]], [[OFFSET]];
|
|
%r4 = call i32 @llvm.nvvm.fns(i32 2, i32 %base, i32 %offset);
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, 2, [[BASE]], 0;
|
|
%r5 = call i32 @llvm.nvvm.fns(i32 2, i32 %base, i32 0);
|
|
%r45 = add i32 %r4, %r5;
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, 2, 1, [[OFFSET]];
|
|
%r6 = call i32 @llvm.nvvm.fns(i32 2, i32 1, i32 %offset);
|
|
; CHECK: fns.b32 {{%r[0-9]+}}, 2, 1, 0;
|
|
%r7 = call i32 @llvm.nvvm.fns(i32 2, i32 1, i32 0);
|
|
%r67 = add i32 %r6, %r7;
|
|
%r4567 = add i32 %r45, %r67;
|
|
%r = add i32 %r0123, %r4567;
|
|
ret i32 %r;
|
|
}
|
|
|