Repository navigation
Retranslate MB03XU; fix ill-posed MB04TB/AB13FD tests - #21
Merged
Merged
Conversation
- mb03xu: 3 of 4 branches were stubs, 4th wrong; blocked MB04TB (ilo>=2, n>~140) gave garbage. Literal retranslation, bitwise match vs Fortran. - mb04tb test: old inputs ill-conditioned + never hit blocked path; replace with ctypes differential (n=200, ilo=2) + blocked vs mb04ts. - ab13fd: info=1 on Accelerate matches Fortran; accept 0/1, add Fortran differential. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
jamestjsp
enabled auto-merge (squash)
October 12, 2026 00:29
…eaf15f # Conflicts: # tests/python/fortran_reference.py
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 02f5b8a141
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Reference MB03XU.f passes LDQ for B(K+I+2,1)/B(1,K+I+2) DGEMVs; wrong when LDB != LDQ (e.g. LTRB, K>0). Add padded-LDB test; pad Q in Fortran differential so reference stays valid. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
Summary
src/MB/mb03xu.c, three of the four transpose branches were placeholder stubs and the fourth was also wrong. Any MB04TB call that took the blocked path (ilo≥2, n ≳ 140) returned wrong results. That also affects mb03xd, ab13dd and ab13hd. Retranslated statement-for-statement fromMB03XU.f, using 1-based access macros. Results now match Fortran exactly in all four transpose combinations.sin/cosdiffer from numpy's in the last bit. It also never reached the blocked path (with ilo=1, nb ≥ nh). Replaced it with a ctypes test that passes Fortran identical inputs (n=200, ilo=2, ldwork=30000, so nb=9), plus a blocked-vs-unblocked (mb04ts) test that needs no Fortran library.reference_cdll()helper intests/python/fortran_reference.py.Test plan
CTRLSYS_FORTRAN_REFERENCE_LIB; the 20 new MB03XU/MB04TB tests fail on the old codeSLICOT-Reference/examples/data(submodule not checked out in this worktree), unrelatedcheck_translation_smells.pyclean🤖 Generated with Claude Code