[Rocprofiler-system]: Fix GPU event enumeration for rocprof-sys-avail and CLI option for parsing GPU HW Counters (#2476)
## Motivation The `rocprof-sys-avail -H -c GPU` command is returning blank output which is expected to display a list of available GPU hardware counters instead. The `rocprof-sys-sample` and `rocprof-sys-run` is missing the `--gpu-events` option for specifying GPU counter events during profiling. ## Technical Details The initialize_event_info() function had a logic bug where it only called set_agents() if the agent_manager was empty, but the actual issue was that the gpu_agents and cpu_agents vectors were empty even when agents were discovered. Fixed the conditional logic to properly call set_agents() when gpu_agents and cpu_agents are empty, regardless of the agent_manager state. Added the `--gpu-events (-G)` option which sets the `ROCPROFSYS_ROCM_EVENTS` environment variable to the specified values. Fixes an issue where unsupported GPU/APU arch is being skipped gracefully - more details about this issue in the below comment.
This commit is contained in:
@@ -519,6 +519,9 @@ main(int argc, char** argv)
|
||||
} catch(std::runtime_error& _e)
|
||||
{
|
||||
verbprintf(0, "Retrieving the GPU HW counters failed: %s", _e.what());
|
||||
} catch(std::exception& _e)
|
||||
{
|
||||
verbprintf(0, "Exception retrieving GPU HW counters: %s", _e.what());
|
||||
}
|
||||
verbprintf(1, "Found %i HIP devices and %zu GPU HW counters\n", gpu_count,
|
||||
_num_metrics);
|
||||
|
||||
Reference in New Issue
Block a user