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

Adding timeout functionality/EnvVar to TestBed (#1044)

Browse Source 
* Adding timeout functionality/EnvVar to TestBed
* updating timeout unit to microseconds

Signed-off-by: Tim Hu <timhu102@amd.com>

[ROCm/rccl commit: 9c0ef11ac7]
This commit is contained in:
Tim
2024-01-17 11:33:01 -05:00
committed by GitHub
parent 1d62a5f440
commit 245e757b26
6 changed files with 53 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rccl/test/common/EnvVars.cpp
+3 -1
View File
@@ -63,6 +63,7 @@ namespace RcclUnitTesting
maxRanksPerGpu = GetEnvVar("UT_MAX_RANKS_PER_GPU", 1);
showTiming = GetEnvVar("UT_SHOW_TIMING", 1);
useInteractive = GetEnvVar("UT_INTERACTIVE", 0);
timeoutUs = GetEnvVar("UT_TIMEOUT_US" , 200000);
// Limit number of supported reduction operators to just ncclSum if only allReduce is built
#ifdef BUILD_ALLREDUCE_ONLY
@@ -195,7 +196,8 @@ namespace RcclUnitTesting
std::make_tuple("UT_MAX_RANKS_PER_GPU", maxRanksPerGpu, "Maximum number of ranks using the same GPU"),
std::make_tuple("UT_PRINT_VALUES" , printValues , "Print array values (-1 for all)"),
std::make_tuple("UT_SHOW_TIMING" , showTiming , "Show timing table"),
std::make_tuple("UT_INTERACTIVE" , useInteractive, "Run in interactive mode")
std::make_tuple("UT_INTERACTIVE" , useInteractive, "Run in interactive mode"),
std::make_tuple("UT_TIMEOUT_US" , timeoutUs , "Timeout limit for collective calls in us")
};
printf("================================================================================\n");