Skip to content

Revert blas + axpy changes #280

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

Merged
merged 2 commits into from
Mar 23, 2017
Merged

Revert blas + axpy changes #280

merged 2 commits into from
Mar 23, 2017

Conversation

bluss
Copy link
Member

@bluss bluss commented Mar 23, 2017

No description provided.

@bluss
Copy link
Member Author

bluss commented Mar 23, 2017

cc @SuperFluffy

@SuperFluffy
Copy link
Contributor

Other than the negative and zero stride case, did you find anything that would prevent using blas for non-contiguous but equispaced/compatible arrays?

@bluss
Copy link
Member Author

bluss commented Mar 23, 2017

It should be fine. Equispaced stride doesn't cover one case though, which is F1 + F2 (both arrays are 2+ dimensions, same shape, but in fortran memory order).

@bluss
Copy link
Member Author

bluss commented Mar 23, 2017

Changed to a revert since it broke existing uses of is_contiguous (for negative stride arrays). It's my responsibility to decide on merges, so it's my responsibility there is a bug.

@bluss
Copy link
Member Author

bluss commented Mar 23, 2017

A clearer thought on the situation is that we have two input arrays, and we need to figure out if they are separately blas compatible, but also that they are memory order compatible with each other. Same problem that zip needs to solve.

@bluss bluss merged commit a53313c into master Mar 23, 2017
@bluss bluss changed the title BUG: Fix blas + axpy by only using it for certain cases Revert blas + axpy changes Mar 23, 2017
@bluss bluss deleted the axpy-fixup branch March 23, 2017 17:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants