Nusrat Islam 9e8cc15c61 Merge pull request #797 from nusislam/allgather-mi250-perf
device: fine tune MI200/MI250 simple protocol performance

[ROCm/rccl commit: c095424e4d]
2023-07-10 09:53:03 -05:00
S
Descrição
Descrição não fornecida
283 MiB
0 Favoritos 2 Observadores 0 Forks
Linguagens
C++ 67.5%
C 20.6%
Python 6.6%
CMake 3.4%
Shell 0.6%
Outra 1.1%