oraluben commented on issue #325:
URL: https://github.com/apache/tvm-ffi/issues/325#issuecomment-3644711877
That's true.
Currently we're migrating an existing kernel library that returns
`at::empty_strided` to runtime so the stride in torch matters. We hope to
minimize the efforts and avoid adding extra `torch.as_strided` calls on python
side, especially when the library is mainly pure C.
Further more, sometime we need to encode additional information in stride:
```
if (a)
result = at::empty_strided({A, B}, {1, A});
else
result = at::empty_strided({A, B}, {B, A});
```
This would be hard to express in current PR.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
