Refactor deepcopy logic to improve planning speed #665
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
facebook-github-bot
added
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D39822605 |
4 similar comments
|
This pull request was exported from Phabricator. Differential Revision: D39822605 |
|
This pull request was exported from Phabricator. Differential Revision: D39822605 |
|
This pull request was exported from Phabricator. Differential Revision: D39822605 |
|
This pull request was exported from Phabricator. Differential Revision: D39822605 |
Summary: Pull Request resolved: #665 # Summary When there are many possible proposals to plan with, e.g. >1000 proposals from GreedyProposer for ads models in [fused, fused_uvm_caching] modes, the planner can run >20 min to come up with a best plan. The speed bottleneck comes from `deepcopy(List[ShardingOption])` in proposers and partitioner. With the diff to reorganize the deepcopy operations, the planning time can be decreased to ~2.5 min. # Idea (1) Decrease times of deepcopy When `List[ShardingOption]` is passed as plan/proposal between proposer, partitioner and planner, two times of `deepcopy(List[ShardingOption])` (one in proposer, the other in partitioner) are done to isolate the state of List[ShardingOption] between planning steps. As deepcopy is expensive, we just need to keep one deepcopy in partitioner. Doing so reduces the planning time from ~20 min to ~10 min. (2) Custom `__deepcopy__` function in ShardingOption Even with just one time of deepcopy in partitioner, the overall planning time is still dominated by that. Looking into the partitioner code, we found the only field updated in ShardingOption is `shard.ranks`. To avoid copying big objects (e.g. tensor, module) to save time, a custom `__deepcopy__` function is created in ShardingOption that only deepcopies everything except `tensors` and `modules`, which are passed with the original object. With this custom deepcopy function, the planning time further decreases from ~10 min to 2.5 min. (3) Freeze `tensor` and `module` in ShardingOption With idea (2), the `tensor` and `module` fields in ShardingOption are changed to be read-only to avoid modification during planning. Reviewed By: bigning Differential Revision: D39822605 fbshipit-source-id: eb99a8efe9ff580a88ebd6fa446059f223337e3a
|
This pull request was exported from Phabricator. Differential Revision: D39822605 |
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.
Summary:
Summary
When there are many possible proposals to plan with, e.g. >1000 proposals from GreedyProposer for ads models in [fused, fused_uvm_caching] modes, the planner can run >20 min to come up with a best plan. The speed bottleneck comes from
deepcopy(List[ShardingOption])in proposers and partitioner. With the diff to reorganize the deepcopy operations, the planning time can be decreased to ~2.5 min.Idea
(1) Reduce times of deepcopy
When
List[ShardingOption]is passed as plan/proposal between proposer, partitioner and planner, two times ofdeepcopy(List[ShardingOption])(one in proposer, the other in partitioner) are done to isolate the state of List[ShardingOption] between planning stesp. As deepcopy is expensive, we just need to keep one deepcopy in partitioner. Doing so reduces the planning time from ~20 min to ~10 min.(2) Custom deepcopy
During partitioning, the only field updated in ShardingOption is
shard.ranks. To avoid copying big objects (e.g. tensor, module) to save time, a custom__deepcopy__function is created in ShardingOption that only deepcopiesshards, and that everything else is passed with the original object. Doing so reduces the planning time from ~10 min to 2.5 min.(3) Freeze _tensor and _module
With idea (2), the
tensorandmodulefields in ShardingOption are changed to be read-only to avoid modification during planning.Differential Revision: D39822605