Skip to content

[TensorRT EP] Address GPU bf16 support check #24915

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
chilo-ms merged 1 commit into from
Jun 1, 2025
Merged

[TensorRT EP] Address GPU bf16 support check #24915

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
5 changes: 5 additions & 0 deletions onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1369,6 +1369,11 @@ TensorrtExecutionProvider::TensorrtExecutionProvider(const TensorrtExecutionProv
max_workspace_size_ = info.max_workspace_size;
fp16_enable_ = info.fp16_enable;
bf16_enable_ = info.bf16_enable;
// BF16 support is primarily available on NVIDIA GPUs with the Ampere and later architectures with compute capability of 8.0 or higher.
if (bf16_enable_ && prop.major < 8) {
bf16_enable_ = false;
LOGS_DEFAULT(WARNING) << "[TensorRT EP] trt_bf16_enable is set, but platform doesn't support bf16.";
}
int8_enable_ = info.int8_enable;
if (int8_enable_) {
int8_calibration_cache_name_ = info.int8_calibration_table_name;
Expand Down
Loading