Skip to content

Route matricize through an ismatricizeview trait - #229

Merged
mtfishman merged 5 commits into
mainfrom
mf/factorization-buffer-reuse
Sep 10, 2026
Merged

Route matricize through an ismatricizeview trait#229
mtfishman merged 5 commits into
mainfrom
mf/factorization-buffer-reuse

Conversation

@mtfishman

Copy link
Copy Markdown
Member

Matricization now declares its memory behavior: an ismatricizeview trait says which splits share storage, matricizeview and matricizecopy implement the two cases, and matricize routes on the trait. Factorizations use the declarations to make exactly one input copy, replacing runtime aliasing inference, and contracting into a swapped dense destination got 4 to 5 times faster.

@codecov

codecov Bot commented Sep 10, 2026

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 76.78571% with 26 lines in your changes missing coverage. Please review.
✅ Project coverage is 74.78%. Comparing base (e2a6df4) to head (4b581bd).

Files with missing lines Patch % Lines
src/factorizations.jl 79.68% 13 Missing ⚠️
src/matricize.jl 70.83% 7 Missing ⚠️
ext/TensorAlgebraTensorKitExt.jl 25.00% 6 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (e2a6df4) and HEAD (4b581bd). Click for more details.

HEAD has 4 uploads less than BASE
Flag BASE (e2a6df4) HEAD (4b581bd)
docs 2 1
6 3
Additional details and impacted files
@@            Coverage Diff             @@
##             main     #229      +/-   ##
==========================================
- Coverage   80.09%   74.78%   -5.32%     
==========================================
  Files          28       27       -1     
  Lines        1050     1047       -3     
==========================================
- Hits          841      783      -58     
- Misses        209      264      +55     
Flag Coverage Δ
docs 22.07% <38.18%> (-0.23%) ⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

mtfishman and others added 2 commits September 10, 2026 17:13
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@mtfishman
mtfishman enabled auto-merge (squash) September 10, 2026 22:46
@mtfishman
mtfishman merged commit a62faa3 into main Sep 10, 2026
15 of 21 checks passed
@mtfishman
mtfishman deleted the mf/factorization-buffer-reuse branch September 10, 2026 22:49
mtfishman added a commit to ITensor/GradedArrays.jl that referenced this pull request Sep 11, 2026
The graded apply path spent most of its time recomputing sector
structure on every call. Now axes, matrices, and contraction outputs
carry it. Gate sweeps run 2 to 3 times faster and match TensorKit.
Builds on ITensor/TensorAlgebra.jl#229.

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
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.

1 participant