Skip to content

Audit readiness to remove frontend mux cleanup - #540

Merged
xtofalex merged 1 commit into
mainfrom
audit/mux-removal-readiness
Nov 16, 2026
Merged

xtofalex merged 1 commit into
mainfrom
audit/mux-removal-readiness

Conversation

@xtofalex

Copy link
Copy Markdown
Collaborator

Both frontend calls to SNLUtils::mergeRepeatedMuxSelections still perform material mux reduction after #535/#536. This report-only audit records direct allocation/pass measurements on main c968cf72, classifies the residual lowering boundaries, and compares an isolated calls-disabled variant without changing production behavior.

Direct cleanup removes 383 muxes / 951 output bits from public Ibex and 18 / 340 from public VHDL AES. Disabling the calls retains those muxes while leaving total instance counts unchanged. Keep both calls enabled. The smallest proposed follow-up is case-local scalar SV replay accumulation; 207 observed case-to-case removals are an opportunity ceiling, not promised coverage. VHDL residuals and utility deletion remain separate assessments. No optimization is implemented here.

Evidence includes repeated folded and hierarchy-expanded counts, per-design pass totals, classified merges, and all 208 uninstrumented timing samples. Timing is exploratory: ranges are wide and controls noisy. All 240 corpus input/source/include hashes match the established public/synthetic configurations. No confidential inputs, raw source copies, diagnostic patches or binary archives are included.

Validation:

  • Eight focused cross-process mux/overwrite determinism tests passed in each uninstrumented variant, covering live state, dumps, snapshots and fresh-process reloads.
  • Diagnostic counts repeat across three fresh processes per workload; enabled diagnostic and uninstrumented final structures agree.
  • reuse lint and git diff --check pass.
  • CMake Release only; no new full frontend/four-state equivalence campaign, Bazel parity or BCR readiness claim.

See the audit report for methods, limitations and the evidence required before removal.

@xtofalex
xtofalex merged commit fcf27c4 into main Nov 16, 2026
33 of 36 checks passed
@xtofalex
xtofalex deleted the audit/mux-removal-readiness branch November 16, 2026 10:39
@codecov-commenter

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@codspeed

codspeed Bot commented Nov 16, 2026

Copy link
Copy Markdown

Merging this PR will not alter performance

1 warning

⚠️ 13 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

✅ 111 untouched benchmarks


Comparing audit/mux-removal-readiness (ecdfa78) with main (c968cf7)

Open in CodSpeed

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.

2 participants