Add NCCL operation for allgather #1952
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
zhouxianhao
force-pushed
the
nccl_allgather
branch
from
db8034a
to
4f00743
Compare
added 2 commits
May 15, 2020 16:46
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
romerojosh
reviewed
May 15, 2020
added 4 commits
May 16, 2020 13:34
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
Signed-off-by: jathonzhou <jathonzhou@tencent.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR is to improve the performance of allgather function when HOROVOD_GPU_ALLGATHER=NCCL. Using ncclAllgather when each rank has the same shape, otherwise using ncclBroadcast. Using ncclAllgather is faster than using ncclBroadcast, and using ncclBroadcast is faster then using MPI_Allgatherv.