Nusrat Islam e1c20e7f24 ext-src: Improved allreduce performance in cpx mode for MI308 (#1393) ...

To get the improved performance for TP=4, the user needs to use
RCCL_MSCCL_FORCE_ENABLE=1 and MSCCLPP_READ_ALLRED=1. For TP=8, the
user should use MSCCLPP_HIERARCHICAL_ALLRED=1.

[ROCm/rccl commit: 0fb3b5eba9]

2024-10-30 08:30:15 -05:00

..

json @ 9cca280a4d

Added nlohmann/json:v3.11.3 as a submodule in ext-src and passed its path into the mscclpp build to avoid downloading the package at build time. (#1330)

2024-09-11 16:54:26 -06:00

mscclpp @ 1e82dd444f

Re-enabled MSCCL++ (#1325)

2024-09-11 09:55:16 -06:00

cpx.patch

Enable MSCCLPP use in CPX mode (#1355)

2024-10-02 11:52:04 -05:00

read-allred.patch

ext-src: Improved allreduce performance in cpx mode for MI308 (#1393)

2024-10-30 08:30:15 -05:00