Skip to content

suppress cuda error messages (harmless) when doing training on ROCm #579

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

Open
cj401-amd wants to merge 3 commits into
base: rocm-jaxlib-v0.8.2
Choose a base branch
from
+30 −0
Open

suppress cuda error messages (harmless) when doing training on ROCm #579

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
f1d3ad7
suppress error (harmless) message related to cuda due to double plugi…
cj401-amd Jan 26, 2026
f72ed81
restore LOG(WARNING) for ComputationPlacer
cj401-amd Jan 26, 2026
be50dc1
update to remove the wrong return
cj401-amd Jan 27, 2026
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
10 changes: 10 additions & 0 deletions xla/stream_executor/cuda/cuda_blas.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1399,6 +1399,16 @@ absl::Status CUDABlas::GetVersion(std::string *version) {
}

void initialize_cublas() {
// Check if already registered before attempting - prevents duplicate
// registration error messages (can happen with multiple library loads)
auto already_registered = PluginRegistry::Instance()->HasFactory(
kCudaPlatformId, PluginKind::kBlas);

if (already_registered) {
// Already registered, skip silently (mimics ROCm behavior)
return;
}

absl::Status status =
PluginRegistry::Instance()->RegisterFactory<PluginRegistry::BlasFactory>(
kCudaPlatformId, "cuBLAS",
Expand Down
10 changes: 10 additions & 0 deletions xla/stream_executor/cuda/cuda_dnn.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6947,6 +6947,16 @@ absl::Status CudnnGraph::PopulateOrUpdateRawCommandBuffer(
} // namespace gpu

void initialize_cudnn() {
// Check if already registered before attempting - prevents duplicate
// registration error messages (can happen with multiple library loads)
auto already_registered = PluginRegistry::Instance()->HasFactory(
cuda::kCudaPlatformId, PluginKind::kDnn);

if (already_registered) {
// Already registered, skip silently (mimics ROCm behavior)
return;
}

absl::Status status =
PluginRegistry::Instance()->RegisterFactory<PluginRegistry::DnnFactory>(
cuda::kCudaPlatformId, "cuDNN",
Expand Down
10 changes: 10 additions & 0 deletions xla/stream_executor/cuda/cuda_fft.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -460,6 +460,16 @@ STREAM_EXECUTOR_CUDA_DEFINE_FFT(double, Z2Z, D2Z, Z2D)
} // namespace gpu

void initialize_cufft() {
// Check if already registered before attempting - prevents duplicate
// registration error messages (can happen with multiple library loads)
auto already_registered = PluginRegistry::Instance()->HasFactory(
cuda::kCudaPlatformId, PluginKind::kFft);

if (already_registered) {
// Already registered, skip silently (mimics ROCm behavior)
return;
}

absl::Status status =
PluginRegistry::Instance()->RegisterFactory<PluginRegistry::FftFactory>(
cuda::kCudaPlatformId, "cuFFT",
Expand Down
Loading