Skip to content

ci: compile the CUDA lane for AMPERE86 too; docs: API fixes need no SEP - #845

Merged
jtravs merged 1 commit into
mainfrom
ci/ampere86-and-sep-rule
Nov 15, 2026
Merged

jtravs merged 1 commit into
mainfrom
ci/ampere86-and-sep-rule

Conversation

@jtravs

@jtravs jtravs commented Nov 15, 2026

Copy link
Copy Markdown
Contributor

Two small maintainer decisions:

CI (follow-up to #815). device-compile-cuda now runs as a matrix over HOPPER90 (the preset's target) and AMPERE86 (NVIDIA A40). The A40 radial crash came from device_dense_gemm, which only an AMPERE86 build compiles, so the lambda-tag check added in #841 could not see it on the old lane. Each CUDA build now runs the check over its own objects.

  • Cost: one more CUDA compile, about 20–28 runner-minutes, on pushes that touch native code. It runs in parallel with the other jobs. The job is renamed device-compile (CUDA 12.8, <arch>); only ci-passed is a required check, and it already depends on this job.
  • docs/dev/tooling.md and the check's docstring describe the new lane.

SEP threshold. AGENTS.md, docs/dev/contributing.md and SEP 0 §2 no longer list "a redesign of a public interface" as needing a SEP. API fixes and cleanups that change no model or physics now follow the ordinary issue-and-PR path, even when they rename, merge or refuse public parameters. Physics defaults and conventions, new models and schemes, and kernel-contract changes still need one. SEP 0's log records the decision with today's date.

This PR changes CI and docs only, with no code behaviour and no changelog fragment.

🤖 Generated with Claude Code

https://claude.ai/code/session_01XGENPwwkUUG9KSAwneHjgH

The device-compile lane now builds the cuda preset for HOPPER90 and for
AMPERE86, and runs the lambda-tag check over each build. Only an AMPERE86
build compiles the device GEMM that caused #815.

AGENTS.md, contributing.md and SEP 0 section 2 no longer list a public
interface redesign as needing a SEP. API fixes and cleanups that change no
model or physics follow the ordinary issue and pull-request path
(maintainer's decision, recorded in SEP 0's log).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XGENPwwkUUG9KSAwneHjgH
@jtravs
jtravs marked this pull request as ready for review November 15, 2026 20:38
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Nov 15, 2026 •

Copy link
Copy Markdown
Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-11-15T20:39:27.672488Z 158c372 Draft marked ready
🔒 Security Review ✅ Completed 2026-11-15T20:40:45.078930Z 158c372 Draft marked ready
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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@read-the-docs-community

Copy link
Copy Markdown

Documentation build overview

📚 solprop | 🛠️ Build #35911644 | 📁 Comparing 158c372 against latest (281a954)

  🔍 Preview build  

3 files changed
± dev/contributing.html
± dev/tooling.html
± dev/seps/sep-0000-process.html

@codecov

codecov Bot commented Nov 15, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@jtravs
jtravs merged commit 33efe0c into main Nov 15, 2026
73 checks passed
@jtravs
jtravs deleted the ci/ampere86-and-sep-rule branch November 15, 2026 22:34
jtravs added a commit that referenced this pull request Nov 16, 2026
The changes job diffed against the event's pull_request.base.sha, which can
lag behind main: it stayed at 281a954 on #847 after #845 merged, so #845's
ci.yml and native/ changes were counted as #847's own and the docs-only PR
ran every lane. The checkout is GitHub's merge of the head into main's current
tip, so its first parent is the right base.

Co-authored-by: Claude Opus 5.5 <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