Physical Design · All levels
Data and Libraries Tricky Q&A
20+ senior Data and Libraries interview questions.
Q&A bank
Answer with mechanism, pitfall, regression check, and release judgment. At 10+ years, a correct definition is not enough.
Senior answer rubric
Start with the failing metric and analysis context.
Explain the physical or constraint mechanism.
Name the cheapest evidence-gathering experiment.
Choose a bounded fix and state what it can regress.
Close with signoff, waiver, or escalation criteria.
What belongs in a technology manifest for PD runs?
[INT][PD]
Q: What belongs in a technology manifest for PD runs?
A:
PDK release tag, deck paths, map files, extraction techfile, tool versions, and checksums.
FOLLOW-UP TRAP: Only listing 'foundry version' without file-level identity.Why can deck compile pass but signoff still be wrong?
[INT][PD]
Q: Why can deck compile pass but signoff still be wrong?
A:
Compile validates syntax/includes; semantic mismatch (wrong layer map or deck release) can still invalidate results.
FOLLOW-UP TRAP: Treating compile success as full signoff equivalence.LEF vs DEF in one sentence?
[INT][PD]
Q: LEF vs DEF in one sentence?
A:
LEF defines reusable abstract physical rules; DEF records design-instance physical state.
FOLLOW-UP TRAP: Saying DEF can replace LEF for library distribution.How do stale macro abstracts usually surface?
[INT][PD]
Q: How do stale macro abstracts usually surface?
A:
Pin access DRC, congestion detours, or parent-route mismatches after IP refresh.
FOLLOW-UP TRAP: Blaming router randomness first.What is a healthy library corner mapping practice?
[INT][PD]
Q: What is a healthy library corner mapping practice?
A:
Explicit setup/hold/leakage mapping per mode with zero unresolved cells and archived manifests.
FOLLOW-UP TRAP: Using whichever lib files load fastest.When should mixed CCS and NLDM be allowed?
[INT][PD]
Q: When should mixed CCS and NLDM be allowed?
A:
Only with explicit methodology qualification and documented impact boundaries.
FOLLOW-UP TRAP: Mixing ad hoc in production without correlation.Why are MMMC views considered contracts?
[INT][PD]
Q: Why are MMMC views considered contracts?
A:
They bind SDC, liberty, RC, and derates; broken bindings change optimization truth.
FOLLOW-UP TRAP: Treating views as optional runtime convenience.What report confirms derates are actually applied?
[INT][PD]
Q: What report confirms derates are actually applied?
A:
Applied derate matrix report per analysis view.
FOLLOW-UP TRAP: Assuming derates loaded because script sourced file.How do you roll out new OCV tables safely?
[INT][PD]
Q: How do you roll out new OCV tables safely?
A:
Side-by-side delta on reference blocks, classify movement, then promote with updated closure criteria.
FOLLOW-UP TRAP: Replacing tables directly on active tapeout branch.What is the biggest risk of unmanaged SDC overrides?
[INT][PD]
Q: What is the biggest risk of unmanaged SDC overrides?
A:
Silent behavior drift between local and CI runs causing non-reproducible closure.
FOLLOW-UP TRAP: Thinking overrides are harmless if WNS improves.How do you gate high-risk SDC changes?
[INT][PD]
Q: How do you gate high-risk SDC changes?
A:
Semantic diff + owner approvals + unconstrained path checks before merge.
FOLLOW-UP TRAP: Reviewing only textual diff volume.What should an IP handoff bundle include beyond LEF?
[INT][PD]
Q: What should an IP handoff bundle include beyond LEF?
A:
Timing model, UPF/power intent, constraints, release notes, and compatibility status.
FOLLOW-UP TRAP: Shipping only LEF and expecting integrator to infer the rest.Why track compatibility as breaking/non-breaking in IP releases?
[INT][PD]
Q: Why track compatibility as breaking/non-breaking in IP releases?
A:
It gives integrators deterministic upgrade expectations and prevents hidden interface regressions.
FOLLOW-UP TRAP: Using one rolling 'latest' tag.How do you detect corner fallback substitutions?
[INT][PD]
Q: How do you detect corner fallback substitutions?
A:
Inspect library resolution logs for unresolved cells and fallback paths.
FOLLOW-UP TRAP: Assuming missing warnings means no substitution.What is a practical first debug step for run-to-run timing swings?
[INT][PD]
Q: What is a practical first debug step for run-to-run timing swings?
A:
Diff resolved libs, SDC tag, and MMMC view bindings before touching optimization settings.
FOLLOW-UP TRAP: Starting with random place-opt seeds.Why pin-contract diff should be automated?
[INT][PD]
Q: Why pin-contract diff should be automated?
A:
Manual review misses subtle geometry/name shifts that break integration late.
FOLLOW-UP TRAP: Trusting release emails as sufficient.How do you prove DEF handoff readiness?
[INT][PD]
Q: How do you prove DEF handoff readiness?
A:
Run strict import in consumer flow and verify zero critical parser/geometry mismatches.
FOLLOW-UP TRAP: Only viewing DEF in producer tool.What is the consequence of outdated extraction techfiles?
[INT][PD]
Q: What is the consequence of outdated extraction techfiles?
A:
RC mismatch between implementation and signoff leading to timing and SI surprises.
FOLLOW-UP TRAP: Assuming extraction only affects parasitic runtime.How should library and constraint releases be linked?
[INT][PD]
Q: How should library and constraint releases be linked?
A:
Both should share a common run manifest and release tag consumed by CI.
FOLLOW-UP TRAP: Versioning them independently without compatibility checks.What makes collateral governance scalable?
[INT][PD]
Q: What makes collateral governance scalable?
A:
Immutable tagged bundles, semantic diffing, machine checks, and explicit ownership.
FOLLOW-UP TRAP: Relying on tribal knowledge and shared folders.How to drill this Q&A bank
Use each question as a two-minute mock. The target answer is not an essay; it is a structured closure response with a metric, mechanism, risk, and follow-up check.
Answer template
MECHANISM: what physical effect or tool step is involved
WHEN: where it appears in the flow
PITFALL: one wrong junior answer
CHECK: report/map/checklist item that proves the answerScoring
5/5: metric, mechanism, experiment, regression, and release decision named without prompting.
4/5: technically correct and names regression, but misses ownership or escalation criteria.
3/5: correct concept but no decision framework.
1/5: tool command or buzzword with no mechanism.
Principal-level review bar
Q&A bank pages in this course should be read like real closure review material. For a 10+ year PD engineer, the bar is not remembering terminology; it is making a release-quality decision under ambiguity.
What excellent looks like
Names the failing metric, corner/mode, database tag, and analysis switches before proposing a fix.
Separates data, constraint, physical, tool, and methodology root causes instead of treating all failures as optimization problems.
Chooses experiments by information gain and reversibility, not by habit.
States regression blast radius across timing, route, power, PV, DFT, package, and tapeout manifest.
Turns recurring failures into methodology guardrails, dashboards, or checklist items.
Closure note template
STAFF / PRINCIPAL CLOSURE NOTE
Context:
stage: <pre-CTS | post-CTS | post-route | post-fill | signoff>
tag: <database / netlist / SDC / library stack>
failing metric: <exact report line>
affected scope: <block / hierarchy / path group / power domain / region>
Hypotheses:
H1: <most likely physical or constraint mechanism>
H2: <competing explanation>
H3: <methodology or input-data issue>
Decision:
next experiment: <cheap check that can falsify H1>
fix candidate: <minimal reversible change>
rollback trigger: <metric that says the fix is wrong>
regression set: <timing / route / power / PV / DFT / package>
escalation owner: <team or reviewer>Tradeoffs a senior engineer must discuss
Technical tradeoff
Interview performance is structured closure thinking under time pressure. Explain not only the preferred fix, but what margin or schedule you are spending to get it.
Cross-team tradeoff
What must RTL, synthesis, CAD, STA, DFT, package, IP, or foundry agree to before this decision is final?
Which artifact becomes the source of truth after the decision: report, waiver, manifest, ECO script, or methodology deck?
What is the cost of being wrong: one rerun, ECO churn, mask risk, performance loss, or silicon escape?
Leadership communication
"The current blocker is <metric> in <corner/mode/stage>. The leading cause is <mechanism>. I recommend <fix> because it is bounded and reversible. The regression surface is <domains>. If it fails, we escalate to <owner> with <evidence>."Key takeaways
Always connect the concept back to a measurable signoff artifact.
A fix is not complete until you can name the regression checks.
Common pitfalls
Optimizing by habit instead of reading the current report.
Forgetting that a local fix can regress timing, routing, power, or PV elsewhere.