-
Notifications
You must be signed in to change notification settings - Fork 689
update cuda delegate resource free pipeline for safety and segfault-free #14905
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/14905
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New FailuresAs of commit ab1c5c1 with merge base f32e9fc ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) ghstack-source-id: 314867937 Pull Request resolved: #14905
This PR needs a
|
// We noticed that AOTInductorModelContainerDelete doesn't work well with | ||
// mutitple .so files when we tried to use it to delete container handle, | ||
// since freeing one of them will free some sharing resources, leading to | ||
// segfault when trying to free the other .so files. Now we do not explicted |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You can let AI proofread these.
// segfault when trying to free the other .so files. Now we do not explicted | |
// segfault when trying to free the other .so files. Now we do not explicitly |
… segfault-free" This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
Pull Request resolved: #14905 This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. ghstack-source-id: 314881464 @exported-using-ghexport Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/)
… segfault-free" This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
Pull Request resolved: #14905 This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. ghstack-source-id: 314887533 @exported-using-ghexport Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/)
… segfault-free" This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
Pull Request resolved: #14905 This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. ghstack-source-id: 314925612 @exported-using-ghexport Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/)
… segfault-free" This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
Pull Request resolved: #14905 This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. ghstack-source-id: 314936353 @exported-using-ghexport Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/)
… segfault-free" This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
Pull Request resolved: #14905 This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. ghstack-source-id: 314948206 @exported-using-ghexport Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/)
… segfault-free" This diff survives `clear_all_tensors()` function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files. Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) [ghstack-poisoned]
0bbcec0
into
gh/gasoonjia/50/base
…ree (#14927) This PR was created by the merge bot to help merge the original PR into the main branch. ghstack PR number: #14905 by @Gasoonjia ^ Please use this as the source of truth for the PR details, comments, and reviews ghstack PR base: https://github.com/pytorch/executorch/tree/gh/gasoonjia/50/base ghstack PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/50/head Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/gasoonjia/49/orig Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/gasoonjia/50/orig Differential Revision: [D84135792](https://our.internmc.facebook.com/intern/diff/D84135792/) @diff-train-skip-merge --------- Co-authored-by: gasoonjia <gasoonjia@icloud.com>
Stack from ghstack (oldest at bottom):
This diff survives
clear_all_tensors()
function and enable it during backend destroy stage. Furthermore, we defer the container handle deletion to OS to avoid potential segfault if there's more than one .so files.Differential Revision: D84135792