AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Enabled profiling for CPU agents for memcpy activities

Browse Source 
To be able to trace memcpy asynchronously, both dst and src agents need to have profiling enabled and the api for enabling profiling was only enabling for gpu agents. CPU agents didn't have profiling enabled so the signal owner could not be known. hsa_amd_profiling_get_async_copy_time will fail with an HSA status error because it can't read the agent for the given signal.

Change-Id: Ie165e0e39b8fcd6992a55695b9ffcead10a8e812


[ROCm/ROCR-Runtime commit: ae1da390bd]
This commit is contained in:
Youssef Aly
2023-11-07 12:56:53 -06:00
committed by David Yat Sin
parent 880ddd4387
commit 1c1298c1c0
1 changed files with 5 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rocr-runtime/runtime/hsa-runtime/core/runtime/hsa_ext_amd.cpp
+5
View File
@@ -414,6 +414,11 @@ hsa_status_t hsa_amd_profiling_async_copy_enable(bool enable) {
hsa_status_t err = agent->profiling_enabled(enable);
if (err != HSA_STATUS_SUCCESS) ret = err;
}
for (core::Agent* agent : core::Runtime::runtime_singleton_->cpu_agents()) {
hsa_status_t err = agent->profiling_enabled(enable);
if (err != HSA_STATUS_SUCCESS) ret = err;
}
return ret;
CATCH;