Use GPUArrays caching allocator in train!#2665
Merged
Merged
Conversation
Use GPUArrays caching allocator in train! Wraps each training iteration in `GPUArrays.@cached` so temporary GPU allocations (gradients, activations) are pooled and reused across steps instead of being freed and reallocated each iteration, reducing GC pressure. Closes #2636 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> cleanup cleanup
CarloLucibello
force-pushed
the
cl/caching-allocator
branch
from
8ab3fa7 to
e4c6bed
Compare
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
train!iteration inGPUArrays.@cachedso temporary GPU allocations (gradients, intermediate activations) are pooled and reused across steps rather than freed and reallocated each timeAllocCachepertrain!call and callsunsafe_free!after the loop for deterministic cleanupGPUArraysas a direct dependency with compat"11.2"(the version that introducedAllocCache)Closes #2636
Test plan
train!tests pass on CPUDomainErroron non-finite loss still works correctly🤖 Generated with Claude Code