[SR] Refactor memory planner to prepare for new algorithm #74730
Conversation
This file contains hidden or 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
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit 4ee901a (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
facebook-github-bot
added
cla signed
oncall: jit
mikeiovine
pushed a commit
that referenced
this pull request
Mar 25, 2022
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! ghstack-source-id: 152185658 Pull Request resolved: #74730
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
This was referenced Mar 29, 2022
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Apr 4, 2022
Pull Request resolved: #74730 Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. ghstack-source-id: 152925429 Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)!
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Apr 5, 2022
Pull Request resolved: #74730 Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. ghstack-source-id: 153052809 Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)!
This was referenced Apr 5, 2022
Closed
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Apr 6, 2022
Pull Request resolved: #74730 Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. ghstack-source-id: 153168603 Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)!
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)! [ghstack-poisoned]
mikeiovine
pushed a commit
that referenced
this pull request
Apr 7, 2022
Pull Request resolved: #74730 Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. ghstack-source-id: 153288210 Differential Revision: [D35132124](https://our.internmc.facebook.com/intern/diff/D35132124/) **NOTE FOR REVIEWERS**: This PR has internal Facebook specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D35132124/)!
facebook-github-bot
pushed a commit
that referenced
this pull request
Apr 7, 2022
Summary: Pull Request resolved: #74730 Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible. My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of `if (use_new_strategy)` statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (like `StorageGroup`). It's much cleaner if we instead turn `MemoryPlanner` into an abstract base class and have different subclasses implement their strategies in `allocateManagedTensors` and `deallocateManagedTensors`. ghstack-source-id: 153288210 Test Plan: Existing unit tests Reviewed By: navahgar, hlu1 Differential Revision: D35132124 fbshipit-source-id: c5ef5ae6361b44dedf97090201e244a76e1e6bce
Contributor
|
Hey @mikeiovine. |
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.
Stack from ghstack (oldest at bottom):
Motivation: I am working on implementing a new, more efficient memory planning algorithm. This algorithm cannot replace the old one entirely, because it can only be practically done for models that have sample inputs to warm up with. We need a way to make the memory planner's strategy extensible.
My first pass attempt at implementing the new algorithm crammed everything into the same class, but it became a nightmare to manage (a ton of
if (use_new_strategy)statements everywhere). Additionally, it was a little clumsy since there are some concepts that make sense for one algorithm but not the other (likeStorageGroup).It's much cleaner if we instead turn
MemoryPlannerinto an abstract base class and have different subclasses implement their strategies inallocateManagedTensorsanddeallocateManagedTensors.Differential Revision: D35132124
NOTE FOR REVIEWERS: This PR has internal Facebook specific changes or comments, please review them on Phabricator!