Skip to content

[OpenMP] Fix or disable NVPTX tests failing currently #77844

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jan 12, 2024
Merged

[OpenMP] Fix or disable NVPTX tests failing currently #77844

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions openmp/libomptarget/test/libc/assert.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,8 @@

// REQUIRES: libc

// NVPTX without LTO uses the implementation in OpenMP currently.
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: powerpc64-ibm-linux-gnu
// UNSUPPORTED: powerpc64-ibm-linux-gnu-LTO
// UNSUPPORTED: aarch64-unknown-linux-gnu
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/mapping/target_derefence_array_pointrs.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,10 @@
// RUN: %libomptarget-run-generic 2>&1 \
// RUN: | %fcheck-generic

// FIXME: This is currently broken on all GPU targets
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <stdio.h>
#include <stdlib.h>
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/mapping/target_uses_allocator.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,9 @@
// RUN: %libomptarget-compile-run-and-check-generic

// FIXME: https://github.com/llvm/llvm-project/issues/77841
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/mapping/target_wrong_use_device_addr.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,9 @@
// RUN: env LIBOMPTARGET_INFO=64 %libomptarget-run-fail-generic 2>&1 \
// RUN: | %fcheck-generic

// FIXME: Fails due to optimized debugging in 'ptxas'
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <stdio.h>

int main() {
Expand Down
7 changes: 5 additions & 2 deletions openmp/libomptarget/test/offloading/bug64959.c
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,8 +1,11 @@
// RUN: %libomptarget-compilexx-run-and-check-generic
// RUN: %libomptarget-compileoptxx-run-and-check-generic
// RUN: %libomptarget-compile-run-and-check-generic
// RUN: %libomptarget-compileopt-run-and-check-generic

// TODO: This requires malloc support for the threads states.
// FIXME: Flaky on all GPU targets.
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/offloading/info.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,9 @@
// RUN: env LIBOMPTARGET_INFO=63 %libomptarget-run-amdgcn-amd-amdhsa 2>&1 | \
// RUN: %fcheck-amdgcn-amd-amdhsa -allow-empty -check-prefixes=INFO,AMDGPU

// FIXME: Fails due to optimized debugging in 'ptxas'.
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>

Expand Down
1 change: 1 addition & 0 deletions openmp/libomptarget/test/offloading/std_complex_arithmetic.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@
// FIXME: This fails to link due to missing math symbols. We should provide the
// needed math functions in the GPU `libm` and require the GPU C library.
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <cassert>
#include <complex>
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/unified_shared_memory/close_enter_exit.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,10 @@
// UNSUPPORTED: clang-6, clang-7, clang-8, clang-9

// Fails on amdgpu with error: GPU Memory Error
// Fails on nvptx with error: an illegal memory access was encountered
// XFAIL: amdgcn-amd-amdhsa
// XFAIL: nvptx64-nvidia-cuda
Copy link
Contributor

@shiltian shiltian Dec 3, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

probably nvptx fixed the issue in some version. this can constantly pass on my local machine with CUDA 12.6.

********************
Unexpectedly Passed Tests (2):
  libomptarget :: nvptx64-nvidia-cuda :: unified_shared_memory/api.c
  libomptarget :: nvptx64-nvidia-cuda :: unified_shared_memory/close_enter_exit.c

// XFAIL: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/unified_shared_memory/close_modifier.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,10 @@
// UNSUPPORTED: clang-6, clang-7, clang-8, clang-9

// amdgpu runtime crash
// Fails on nvptx with error: an illegal memory access was encountered
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>
Expand Down
3 changes: 3 additions & 0 deletions openmp/libomptarget/test/unified_shared_memory/shared_update.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,10 @@
// REQUIRES: unified_shared_memory

// amdgpu runtime crash
// Fails on nvptx with error: an illegal memory access was encountered
// UNSUPPORTED: amdgcn-amd-amdhsa
// UNSUPPORTED: nvptx64-nvidia-cuda
// UNSUPPORTED: nvptx64-nvidia-cuda-LTO

#include <omp.h>
#include <stdio.h>
Expand Down