24016346863
91%

Ran 06 Apr 2026 02:44AM UTC

Jobs 1

Files 49

Run time 1min

Badge

Embed ▾

Committed 06 Apr 2026 02:32AM UTC coverage: 91.174%. Remained the same

Build # 24016346863

Build Type

push

github

Committed by

web-flow

Commit Message

[FIX] More realistic compiled memory measurement (#289)

* [FIX] Remove misleading peak memory reset in compiled benchmarks

The compiled memory measurement did a warmup call, then
reset_peak_memory_stats(), then measured only the compiled matvec.
This excluded construction and compilation memory, producing
artificially low numbers compared to the eager path which measured
the full pipeline.

Remove the reset so bench.memory() captures the full pipeline
(setup + compilation + matvec) — what a user would actually experience.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* [DATA] Re-run compiled peak memory benchmarks after fix

Compiled peak memory now includes the full pipeline (setup +
compilation + matvec) instead of only the steady-state matvec.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* [DOC] Add changelog entry for compiled memory fix

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* [FIX] Remove stale peakmem_compiled from EKFAC inverse (fx) nanoGPT

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Felix Dangel <fdangel@vectorinstitute.ai>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Coverage Stats

2407 of 2640 relevant lines covered (91.17%)

0.91 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
1	24016346863.1	06 Apr 2026 02:44AM UTC	49	91.17	GitHub Action Run

f-dangel / curvlinops / 24016346863
91%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 24016346863

f-dangel / curvlinops / 24016346863 91%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 24016346863

f-dangel / curvlinops / 24016346863
91%

README BADGES
x