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

Implement the stats subsystem in rdci

Browse Source 
Add support for the stats subsystem in rdci
Modify the dmon system to handle the case when no GPUs in a group

Change-Id: I5a18e1201d24b5318b8e324a77551a757b108f25
This commit is contained in:
Bill(Shuzhou) Liu
2020-04-21 09:48:14 -04:00
committed by Chris Freehill
parent 1b58033183
commit 096dc2dadb
7 changed files with 369 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rdci/src/RdciDmonSubSystem.cc
+22 -2
View File
@@ -309,12 +309,32 @@ void RdciDmonSubSystem::process() {
result = rdc_group_gpu_get_info(rdc_handle_,
options_[OPTIONS_GROUP_ID], &group_info);
if (result != RDC_ST_OK) {
throw RdcException(result, rdc_status_string(result));
std::string error_msg = rdc_status_string(result);
if (result == RDC_ST_NOT_FOUND) {
error_msg = "Cannot find the group " +
std::to_string(options_[OPTIONS_GROUP_ID]);
}
throw RdcException(result, error_msg.c_str());
}
if (group_info.count == 0) {
throw RdcException(RDC_ST_NOT_FOUND, "The gpu group " +
std::to_string(options_[OPTIONS_GROUP_ID])
+ " must contain at least 1 GPU.");
}
result = rdc_group_field_get_info(rdc_handle_,
options_[OPTIONS_FIELD_GROUP_ID], &field_info);
if (result != RDC_ST_OK) {
throw RdcException(result, rdc_status_string(result));
std::string error_msg = rdc_status_string(result);
if (result == RDC_ST_NOT_FOUND) {
error_msg = "Cannot find the field group " +
std::to_string(options_[OPTIONS_FIELD_GROUP_ID]);
}
throw RdcException(result, error_msg.c_str());
}
if (field_info.count == 0) {
throw RdcException(RDC_ST_NOT_FOUND, "The field group " +
std::to_string(options_[OPTIONS_FIELD_GROUP_ID])
+ " must contain at least 1 field.");
}
// keep extra 1 minute data