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
58 lines
2.1 KiB
LLVM
58 lines
2.1 KiB
LLVM
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_20 | FileCheck %s
|
|
; RUN: %if ptxas %{ llc < %s -mtriple=nvptx64 -mcpu=sm_20 | %ptxas-verify %}
|
|
|
|
; Checks how NVPTX lowers alloca buffers and their passing to functions.
|
|
;
|
|
; Produced with the following CUDA code:
|
|
; extern "C" __attribute__((device)) void callee(ptr f, char* buf);
|
|
;
|
|
; extern "C" __attribute__((global)) void kernel_func(ptr a) {
|
|
; char buf[4 * sizeof(float)];
|
|
; *(reinterpret_cast<ptr>(&buf[0])) = a[0];
|
|
; *(reinterpret_cast<ptr>(&buf[1])) = a[1];
|
|
; *(reinterpret_cast<ptr>(&buf[2])) = a[2];
|
|
; *(reinterpret_cast<ptr>(&buf[3])) = a[3];
|
|
; callee(a, buf);
|
|
; }
|
|
|
|
; CHECK: .visible .entry kernel_func
|
|
define ptx_kernel void @kernel_func(ptr %a) {
|
|
entry:
|
|
%buf = alloca [16 x i8], align 4
|
|
|
|
; CHECK: .local .align 4 .b8 __local_depot0[16]
|
|
; CHECK: mov.b64 %SPL
|
|
|
|
; CHECK: ld.param.b64 %rd[[A_REG:[0-9]+]], [kernel_func_param_0]
|
|
; CHECK: cvta.to.global.u64 %rd[[A1_REG:[0-9]+]], %rd[[A_REG]]
|
|
; CHECK: add.u64 %rd[[SP_REG:[0-9]+]], %SP, 0
|
|
; CHECK: ld.global.b32 %r[[A0_REG:[0-9]+]], [%rd[[A1_REG]]]
|
|
; CHECK: st.local.b32 [{{%rd[0-9]+}}], %r[[A0_REG]]
|
|
|
|
%0 = load float, ptr %a, align 4
|
|
store float %0, ptr %buf, align 4
|
|
%arrayidx2 = getelementptr inbounds float, ptr %a, i64 1
|
|
%1 = load float, ptr %arrayidx2, align 4
|
|
%arrayidx3 = getelementptr inbounds [16 x i8], ptr %buf, i64 0, i64 1
|
|
store float %1, ptr %arrayidx3, align 4
|
|
%arrayidx4 = getelementptr inbounds float, ptr %a, i64 2
|
|
%2 = load float, ptr %arrayidx4, align 4
|
|
%arrayidx5 = getelementptr inbounds [16 x i8], ptr %buf, i64 0, i64 2
|
|
store float %2, ptr %arrayidx5, align 4
|
|
%arrayidx6 = getelementptr inbounds float, ptr %a, i64 3
|
|
%3 = load float, ptr %arrayidx6, align 4
|
|
%arrayidx7 = getelementptr inbounds [16 x i8], ptr %buf, i64 0, i64 3
|
|
store float %3, ptr %arrayidx7, align 4
|
|
|
|
; CHECK-DAG: .param .b64 param0;
|
|
; CHECK-DAG: .param .b64 param1;
|
|
; CHECK-DAG: st.param.b64 [param0], %rd[[A_REG]]
|
|
; CHECK-DAG: st.param.b64 [param1], %rd[[SP_REG]]
|
|
; CHECK: call.uni callee,
|
|
|
|
call void @callee(ptr %a, ptr %buf) #2
|
|
ret void
|
|
}
|
|
|
|
declare void @callee(ptr, ptr)
|