File tree Expand file tree Collapse file tree 1 file changed +4
-4
lines changed
docs/software/communication Expand file tree Collapse file tree 1 file changed +4
-4
lines changed Original file line number Diff line number Diff line change 22# Slingshot network. If the plugin can not be found, applications will fail to
33# start. With the default value, applications would instead fall back to e.g.
44# TCP, which would be significantly slower than with the plugin. More information
5- # about `NCCL_NET` can be found at
6- # https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/env.html#nccl-net.
5+ # about `NCCL_NET` can be found at:
6+ # https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/env.html#nccl-net
77export NCCL_NET="AWS Libfabric"
88# Use GPU Direct RDMA when GPU and NIC are on the same NUMA node. More
9- # information about `NCCL_NET_GDR_LEVEL` can be found at
10- # https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/env.html#nccl-net-gdr-level-formerly-nccl-ib-gdr-level.
9+ # information about `NCCL_NET_GDR_LEVEL` can be found at:
10+ # https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/env.html#nccl-net-gdr-level-formerly-nccl-ib-gdr-level
1111export NCCL_NET_GDR_LEVEL=PHB
1212export NCCL_CROSS_NIC=1
1313# These `FI` (libfabric) environment variables have been found to give the best
You can’t perform that action at this time.
0 commit comments