The profiling tool with Triton Inference Server

The profiling tool with Triton Inference Server

Hi experts,

We would like to monitor the following status on A100 with Triton Inference Server.

  • TensorCore is working or not
  • The whole execution time
  • The GPU execution time
  • Memory utilization

Do you have tools to monitor these status?
We know DLProf can not be used on Triton, so we’re looking for the solution.

Best Regards,
NY

Perf Analyzer can help with some of the things you mentioned.

tgerdes-san,

Thank you for your reply.
We’ll check it.

Best Regards,
NY

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.