ECO Regression Matrix: Expanded Case Study

Extended case study

Signoff review: corners re-run, new violations introduced, signoff delta fails after an apparently minor change affecting ECO Regression Matrix.

Background

Team had closure last week on full MMMC matrix. New run shows regression clustered in one path group.

Symptoms observed

• corners re-run, new violations introduced, signoff delta regression

• Disagreement between synth and STA WNS

• Hold violations after setup ECO

Investigation timeline

1. Hour 0: freeze netlist, SDC, SPEF, lib tags

2. Hour 1: diff constraints and run manifest

3. Hour 2: isolate corner/mode that fails

4. Hour 3: read worst path — cell vs net, clocks

5. Hour 4: classify root cause

6. Hour 5: minimal ECO or SDC fix

7. Hour 6: full regression + signoff memo

Root cause

Root cause tied to ECO Regression Matrix: Every ECO must re-run the agreed MMMC matrix plus SI; regression hygiene prevents shipping a fix that breaks another corner.

Fix and validation

• Bounded ECO or audited SDC change

• Re-run regression manifest, corner comparison table, signoff approval

• Full MMMC + SI regression

Lessons learned

• Tag every run

• Diff constraints before blaming PD

• Setup and hold together

CASE STUDY — ECO Regression Matrix
baseline WNS / regressed WNS / post-fix WNS

Sequence under stress

TIMING PATH — ECO Regression Matrix

launch FF --logic stages--> capture FF
     |                           |
  launch clock                capture clock
     |                           |
  required: arrival < capture_edge - setup
  metric: corners re-run, new violations introduced, signoff delta

STA deep dive

ECO without regression is a coin flip.

Concept diagram

ECO LOOP

failing path -> fix -> setup check -> hold check -> SI -> MMMC matrix

Metric graph

ECO PROGRESS

WNS
 0 |              o target
-50|        o
-100|   o baseline
    +-----------------> ECO #

Reports and artifacts

• pre/post WNS

• hold delta

• leakage delta

• ECO ticket log

Mini case study

48-hour tapeout crunch: VT swap fixed setup; hold fixed with route on short path — order mattered.

Debug branches

• Never ship setup-only proof

• Logical ECO needs equivalence

Senior review question

Ask: what corner/mode/view proves this topic is closed or failing?

Key takeaways

• State corner, mode, view, and database tag with every slack claim.

• Run setup and hold plus MMMC regression after every ECO.

Common pitfalls

• Comparing STA runs with different SPEF or SDC tags.

• Broad false_path to green-wash violations.

• Setup-only ECO without hold check.

Principal STA review addendum

Every ECO must re-run the agreed MMMC matrix plus SI; regression hygiene prevents shipping a fix that breaks another corner.

Metric: corners re-run, new violations introduced, signoff delta