fix is_scalar_on_cpu bug #866
Merged
Conversation
This file contains 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
zhaoguochun1995
requested review from
mrdanielw,
fandaoyi and
lljbash
as code owners
zhaoguochun1995
force-pushed
the
zgc/dipu_fix_is_scalar_on_cpu_bug
branch
from
59e7112
to
216e244
Compare
yewentao256
approved these changes
Jul 4, 2024
caikun-pjlab
approved these changes
Jul 4, 2024
If the tensor on the CPU has only one element and its shape is not empty, such as [1], [1,1], [1,1,...], it should be treated as a Scalar and the corresponding kernel function should be called 而且并没有没对齐，pt 里面 numel 也是返回的1 而不是0 |
lljbash
approved these changes
Jul 9, 2024
yangbofun
approved these changes
Jul 9, 2024
Wrench-Git
pushed a commit
to DeepLink-org/deeplink.framework.dev
that referenced
this pull request
Jul 10, 2024
* fix is_scalar_on_cpu bug * Use is_scalar_on_cpu inline function instead of direct judgment
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Outdated suggestions cannot be applied.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.
Avoid treating the tensor on the device as a CPU scalar tensor. syncStream will be called internally by tensor.item(), which is costly. If the tensor on the CPU has only one element and its shape is not empty, such as [1], [1,1], [1,1,...], it should be treated as a Scalar and the corresponding kernel function should be called. There are places in core where a scalar is wrapped but not marked as wrapped.
see https://github.com/pytorch/pytorch/blob/8f70bf7a943799b5cd870952d39f36361de4b87f/torch/csrc/lazy/core/tensor.cpp#L386