Conversation
**Root cause:** `ForkingPickler` calls `storage._share_cuda_()` on GPU tensors, creating permanent entries in the CUDA IPC cache that hold strong references to GPU memory. These entries are only released when `torch.cuda.ipc_collect()` detects the consumer has closed its IPC handle. **Fix (in `update_weight_from_tensor.py`):** 1. `del hf_named_tensors` added alongside `long_lived_tensors` to break chunk overlap 2. `torch.cuda.ipc_collect()` after each chunk's `ray.get()` + `del` — releases IPC cache entries for completed chunks 3. `torch.cuda.ipc_collect()` after the post-loop barrier — releases the last chunk's IPC entries for non-source ranks (which don't wait for `ray.get()`)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Root cause:
ForkingPicklercallsstorage._share_cuda_()on GPU tensors, creating permanent entries in the CUDA IPC cache that hold strong references to GPU memory. These entries are only released whentorch.cuda.ipc_collect()detects the consumer has closed its IPC handle.Fix (in
update_weight_from_tensor.py):del hf_named_tensorsadded alongsidelong_lived_tensorsto break chunk overlaptorch.cuda.ipc_collect()after each chunk'sray.get()+del— releases IPC cache entries for completed chunkstorch.cuda.ipc_collect()after the post-loop barrier — releases the last chunk's IPC entries for non-source ranks (which don't wait forray.get())