ci: compile both production tiers (x86-64-v4 and sapphirerapids)

[AdaWorldAPI]

Why

Production runs on x86-64-v4 or better, and Railway hosts are Sapphire Rapids, Emerald Rapids or newer. lance-graph's canonical deploy file (Dockerfile.avx512, AdaWorldAPI/lance-graph#1529) now builds sapphirerapids. Until now ndarray's CI compiled only the v4 arm, so the code that sapphirerapids turns on (VNNI, BF16, FP16) was never compiled in CI.

Change

• tier4-avx512-check in ci.yaml becomes a two-entry matrix: x86-64-v4 (.cargo/config-v4.toml) and sapphirerapids (.cargo/config-avx512.toml). Each entry runs both cargo check feature sets and the asm-only codegen witness. Nothing is executed; GitHub's runners are mostly v3.
• .cargo/config-avx512.toml gains -Dwarnings, matching config-v4.toml. A second quoted --config would not work instead: codegen-witness.sh word-splits CARGO_ARGS, so the quotes would reach cargo literally.
• The job sets CARGO_PROFILE_DEV_DEBUG=0, CARGO_PROFILE_TEST_DEBUG=0 and CARGO_INCREMENTAL=0. With --target, build scripts no longer get RUSTFLAGS-style flags and take the profile's debug level instead.
• Each entry has its own rust-cache key.

Measured (local, Cascade Lake host, so nothing was run)

• cargo check --target x86_64-unknown-linux-gnu at sapphirerapids with -Dwarnings passes for approx,serde,rayon and approx,serde,rayon,hpc-extras.
• cargo check -v shows ndarray's rustc call receiving -Ctarget-cpu=native, then -Ctarget-cpu=sapphirerapids (the later flag wins), then -Dwarnings.
• WITNESS_NO_RUN=1 CARGO_ARGS="--config .cargo/config-avx512.toml" scripts/codegen-witness.sh avx512: PASS (1, 1 and 6 vpternlog).

Open

• The job's env -u RUSTFLAGS drops the global crypto cfgs (dalek, poly1305), as it did before; unchanged here.

Blackboard updated in the same commit.

🤖 Generated with Claude Code

https://claude.ai/code/session_011BMJnDBAfzcuQh5HeT39Rh

---

Generated by Claude Code

[coderabbitai]

Warning

Review limit reached

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

Next included review available in 43 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available. Your 62 included PR review attempts over the past 7 days set your current allowance at 1 review per hour.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Essentials
• Run ID: 08819582-c468-47da-b3bc-ea1d6ad6eba0

📥 Commits

Reviewing files that changed from the base of the PR and between 21948f8 and 87ee47b.

📒 Files selected for processing (3)

• .cargo/config-avx512.toml
• .claude/blackboard.md
• .github/workflows/ci.yaml

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Warning

Billing warning: we have not been able to collect payment for this subscription for more than 72 hours. Please update the payment method or pay any pending invoices in Billing to avoid service interruption.

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

Codex Review Summary

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

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-13T06:26:50.715317Z	6b993c0	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.