Skip to content

Conversation

@kshyatt
Copy link
Member

@kshyatt kshyatt commented Jan 8, 2026

Can go in before or after #148, but not having a new version is holding back the CUDA factorizations at TensorKit

@kshyatt kshyatt requested a review from lkdvos January 8, 2026 11:46
@kshyatt
Copy link
Member Author

kshyatt commented Jan 8, 2026

Draft release notes:

  • Truncation for eigensolvers and SVD now supported for GPU arrays
  • Polar decompositions now guarantee results are hermitian
  • Moved orthnull and SVD tests into testsuite
  • Small fixes and improvements to YALAPACK and docs

@lkdvos
Copy link
Member

lkdvos commented Jan 8, 2026

Can we also update the changelog (and try to keep that more updated throughout our PRs?) I really think that is quite important... (I usually start from an automatically generated version with some LLM, and then manually correct)

@kshyatt
Copy link
Member Author

kshyatt commented Jan 8, 2026

Sure will do.

@lkdvos
Copy link
Member

lkdvos commented Jan 8, 2026

I'd add to the release notes:

  • eigenvalue decompositions have consistent sorting for diagonal inputs

@lkdvos
Copy link
Member

lkdvos commented Jan 8, 2026

Since this will fail due to the 1.12 GPU breaking, I will merge this as-is

@lkdvos lkdvos merged commit 91a8a69 into main Jan 8, 2026
7 of 8 checks passed
@lkdvos lkdvos deleted the ksh/bump branch January 8, 2026 16:02
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.

3 participants