fix: stop the WASM engine counting bus accesses on top of per-instruction cycles

[stid]

Closes #215

Summary

• read_byte_from_bus / write_byte_to_bus in wasm-cpu/src/cpu.rs each added a cycle while every opcode arm also added the instruction's full documented count, so step_with_bus reported ~2× (NOP=3, LDA #=4, LDA zp=6, LDA (zp),Y=10). Clock budgets against that value, so the WASM engine executed about half the instructions per emulated second that the JS engine did.
• Fix: drop the two per-access increments (the issue's option 2). The arms alone now match the datasheet and the JS engine's model; IRQ/NMI stay a flat 7 in both engines.
• Tests (TDD, committed red first): compareEngines now asserts the cycle count returned by each performSingleStep matches, and a new documented-cycle battery holds both engines to the datasheet (NOP, LDA imm/zp/abs, abs,X and (zp),Y with and without page cross, STA abs,X, INC abs, JSR/RTS, branch not-taken/taken/page-cross, PHA/PLA). Before the fix: 23/27 red on cycles only.
• Docs: the wasm-performance.md "~14×" headline was measured with the inflated counter. Re-ran BENCH=1 with honest counting → WASM 443 M vs JS 56 M cycles/s = ~7.8× (instr/s, never affected, says ~8.5×). Table + CLAUDE.md updated.
• Version 4.51.9 → 4.51.12 (fix: add missing spacing.xxs token so py-xxs badges get their padding #216 and chore: refresh dependencies (npm minor/patch, Cargo.lock, GitHub Actions) #217 take .10/.11; whichever merges later needs a one-line rebase of src/version.ts).

Verification against real WASM (not just cargo check)

Run with yarn dev:vite serving the rebuilt release build, per cpu_test_guidelines.md:

• engine-parity suite: 27/27 (was 23 red)
• all src/core/cpu-engines/__tests__ suites: 75 passed, 1 skipped (benchmark gate)
• yarn test:ci with the server up: 818 passed, 1 skipped — the parity suites ran for real inside the gate
• cargo check clean; yarn wasm:test shows only the documented wasm-bindgen native-target panics (pre-existing)

LCD: Standard lane, JOURNAL at docs/lcd/work/wasm-cycle-double-count/.

🤖 Generated with Claude Code

https://claude.ai/code/session_01VDq6at6VpsyBxLqB8wBnDT

[coderabbitai]

Warning

Review limit reached

Next included review available in 12 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Team

Run ID: 39761a1e-3f90-4a49-a6e4-a796754652ca

📥 Commits

Reviewing files that changed from the base of the PR and between fc58be0 and 53f8f24.

📒 Files selected for processing (7)

• CLAUDE.md
• docs/active/wasm-performance.md
• docs/lcd/triage-log.md
• docs/lcd/work/wasm-cycle-double-count/JOURNAL.md
• src/core/cpu-engines/__tests__/engine-parity.vitest.test.ts
• src/version.ts
• wasm-cpu/src/cpu.rs

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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