Add support for additional reduction operations for allreduce (min, max, product). #3660
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
Unit Test Results (with flaky tests) 1 201 files - 3 1 201 suites - 3 11h 32m 32s ⏱️ - 10m 6s Results for commit 1129880. ± Comparison against base commit 25ed803. ♻️ This comment has been updated with latest results. |
romerojosh
force-pushed
the
add_allreduce_ops
branch
from
42edd02
to
65d51c9
Compare
…ax, product). Signed-off-by: Josh Romero <joshr@nvidia.com>
Signed-off-by: Josh Romero <joshr@nvidia.com>
Signed-off-by: Josh Romero <joshr@nvidia.com>
Signed-off-by: Josh Romero <joshr@nvidia.com>
Signed-off-by: Josh Romero <joshr@nvidia.com>
Signed-off-by: Josh Romero <joshr@nvidia.com>
romerojosh
force-pushed
the
add_allreduce_ops
branch
from
65d51c9
to
9c26bf8
Compare
maxhgerlach
reviewed
Sep 14, 2022
There was a problem hiding this comment.
I like this PR a lot, nice work @romerojosh! Not only does it add useful functionality, multiple code paths are cleaner now thanks to effectively using the ReduceOp abstraction.
I've left some minor comments.
| @@ -490,7 +490,7 @@ void DoHorovodOperationCudaOnCPU(void*, void* on_complete_ptr, void* param) { | |||
| enqueue_result = EnqueueTensorAllreduces( | |||
| hvd_contexts, hvd_cpu_buffers, hvd_cpu_buffers, ready_event_lists, | |||
| ops_param->op_names, device, callbacks, | |||
| (average) ? ReduceOp::AVERAGE : ReduceOp::SUM, prescale_factor, | |||
| reduce_op, prescale_factor, | |||
There was a problem hiding this comment.
great that the extra layers of converting between average and reduce_op arguments in MXNet could be removed now!
Signed-off-by: Josh Romero <joshr@nvidia.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.
Checklist before submitting
Description
Currently, Horovod only supports allreduce operations using sum or average reduction operations. This PR implements an expanded set of supported reduce operations (min, max, product) for allreduce.
This PR includes changes from #3646 so should not be merged until that one lands.