Nusrat Islam aa8a5f3b2a device: fine tune MI200/MI250 simple protocol performance ...

With Simple protocol, unroll factor of 4 offers better
performance for most of the collectives (on MI200. MI250, and
MI300) except large message allreduce with Ring algorithm
on MI250 and MI200). This PR changes the default unroll factor
to 4 while adding fine tuning for reduction operations.


[ROCm/rccl commit: 58e53dfd37]

2023-07-08 20:21:18 -05:00

projects/rccl

device: fine tune MI200/MI250 simple protocol performance

2023-07-08 20:21:18 -05:00

S

Описание

No description provided

282 MiB

Languages

C++ 67.5%

C 20.6%

Python 6.6%

CMake 3.4%

Shell 0.6%

Разное 1.1%