microsoft/mscclpp
1
0
Fork 0
mirror of https://github.com/microsoft/mscclpp.git synced 2026-04-19 22:39:11 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix multi-nodes test failure (#262)

Browse Source 
fix multi-nodes CI pipeline

Co-authored-by: Changho Hwang <changhohwang@microsoft.com>
This commit is contained in:
Binyang Li
2024-02-07 18:21:05 -08:00
committed by GitHub
parent 2101f5251e
commit 7c229fbdd8
1 changed files with 1 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
python/mscclpp_benchmark/allreduce_bench.py
View File

@@ -165,9 +165,8 @@ def run_benchmark(
memory_out = cp.zeros(nelem, dtype=data_type)
cp.cuda.runtime.deviceSynchronize()
proxy_service = None
proxy_service = ProxyService()
if MPI.COMM_WORLD.size // N_GPUS_PER_NODE == 1:
proxy_service = ProxyService()
if memory.nbytes < 2**20:
mscclpp_algos = [MscclppAllReduce2(mscclpp_group, memory, memory_out)]
else: