Add support for inter-node communication using sockets and InfiniBand/RoCE. Improve latency. Add support for aggregation. Improve LL/regular tuning. Remove tests as those are now at github.com/nvidia/nccl-tests . [ROCm/rccl commit: f93fe9bfd9]
f93fe9bfd9
Change-Id: I92d32b15e516a39710b676aee692ae9b70638937 Reviewed-on: http://git-master/r/935458 Reviewed-by: Przemek Tredak <ptredak@nvidia.com> Tested-by: Przemek Tredak <ptredak@nvidia.com> [ROCm/rccl commit: 130ee246e2]
130ee246e2
[ROCm/rccl commit: 0673d5f44f]
0673d5f44f