Avoid complex-to-real cast warning in CopyBackward #60021
Closed
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
Dropping the imaginary component is expected and gives the correct gradient formula, so silencing the warning is appropriate. [ghstack-poisoned]
💊 CI failures summary and remediationsAs of commit 3b16d43 (more details on the Dr. CI page and at hud.pytorch.org/pr/60021):
🕵️ 2 new failures recognized by patternsThe following CI failures do not appear to be due to upstream breakages:
|
Dropping the imaginary component is expected and gives the correct gradient formula, so silencing the warning is appropriate. [ghstack-poisoned]
peterbell10
added
the
module: complex
albanD
reviewed
Jun 15, 2021
| const bool copy = (grad->is_cuda() && grad->device() != src_device); | ||
| grad_inputs[1] = grad->to( | ||
| src_options, | ||
| /*non_blocking=*/false, |
There was a problem hiding this comment.
Why is the non_blockingness of the forward not propagated to here?
There was a problem hiding this comment.
I think that creates lifetime issues. grad needs to be known to be alive at least until the next device sync, which I don't think we can guarantee.
There was a problem hiding this comment.
cc @mcarilli we might want to double check that one later as part of the syncing in the backward.
|
@mruberry ping |
|
@mruberry has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator. |
zou3519
reviewed
Jul 9, 2021
| } else { | ||
| grad_inputs[1] = grad.to(src_options); | ||
| } | ||
| const bool copy = (grad->is_cuda() && grad->device() != src_device); |
There was a problem hiding this comment.
Do we need to compute copy at all? If grad's device is different from the src device, then it shouldn't matter if we pass grad->to a copy parameter. Because the devices are always different, Tensor::to will always copy regardless of what the copy parameter is. Am I missing something?
There was a problem hiding this comment.
That's what the next PR in the stack changes :)
There was a problem hiding this comment.
Aha! Thanks for the clarification, I should have read the stack :)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
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:
Dropping the imaginary component is expected and gives the correct gradient
formula, so silencing the warning is appropriate.
Differential Revision: D29589371