BMC vs Unbounded Proof Strategies

What this topic teaches

BMC vs Unbounded Proof Strategies converts formal concepts into release-ready verification decisions. Bounded model checking (BMC) searches for counterexamples up to depth k and is excellent for quickly finding shallow bugs, initialization escapes, and protocol startup issues. Unbounded proof attempts establish correctness for all time, typically through induction, interpolation, IC3/PDR-style fixed-point reasoning, or hybrid engine orchestration. K-induction bridges these worlds by proving a base case and inductive step, but it often needs strengthening invariants and helper assertions before convergence. Teams should classify properties early into bug-hunting, bounded-signoff, or unbounded-signoff intent so runtime budgets and expectations stay realistic.

Senior-engineer framing question

When Bound depth achieved, k-induction success rate, and ratio of bug-find to full-proof properties. regresses, can you isolate the first failing assumption/property boundary, prove causality, assign owner, and close with auditable risk?

FORMAL EXECUTION FLOW - BMC vs Unbounded Proof Strategies

requirement intent and risk class
      |
      v
property and assumption modeling
      |
      v
proof engine exploration and trace extraction
      |
      v
counterexample classification and fix hypothesis
      |
      v
re-proof, coverage audit, and signoff decision

Evidence to collect

• Primary metric: Bound depth achieved, k-induction success rate, and ratio of bug-find to full-proof properties..

• Primary artifact: Proof intent matrix mapping each property to BMC depth goals, unbounded targets, and escalation criteria..

• Owners to include: formal verification owner, microarchitecture owner, verification lead, quality/signoff owner.

• One reproducible failing trace and one stable comparator run.

• One fixed metadata run with assumptions and tool settings locked.

Ownership layers

OWNERSHIP LAYERS - BMC vs Unbounded Proof Strategies

+----------------------+--------------------------------+--------------------------------+
| Team                 | Primary responsibility         | Closure artifact               |
+----------------------+--------------------------------+--------------------------------+
| formal verification owner | property and model integrity      | assumptions and proof packet   |
| microarchitecture owner | implementation root-cause closure | RTL fix and replay evidence    |
| verification lead | signoff governance and rollout    | risk memo + acceptance gates   |
+----------------------+--------------------------------+--------------------------------+

Decision matrix

EVIDENCE MATRIX - BMC vs Unbounded Proof Strategies

+-----------------------------+--------------------------------+--------------------------------+---------------------------+
| Evidence                    | Tells you                      | Does not prove                 | Next action               |
+-----------------------------+--------------------------------+--------------------------------+---------------------------+
| property status by class    | closure shape by requirement   | model realism                  | pair with cover reachability |
| vacuity and trigger checks  | assertion meaningfulness       | full legal-path exploration    | inspect assumptions       |
| counterexample traces       | concrete divergence path       | complete bug-space closure     | classify and replay       |
| assumption audit trail      | model boundary confidence      | implementation correctness     | review spec traceability  |
| before/after trend packet   | mitigation movement quality    | long-window stability          | run broader matrix        |
+-----------------------------+--------------------------------+--------------------------------+---------------------------+

Key takeaways

• Start with first-divergence classification before broad model edits.

• Tie each claim to one proving artifact and one owner action.

• Close with residual-risk statement and rollback-safe criteria.

Common pitfalls

• Treating green status as correctness without vacuity and reachability audits.

• Changing assumptions and RTL together, destroying causality.

• Declaring closure without replaying representative legal scenarios.

Formal deep dive

Convergence requires engine strategy, invariant quality, and model realism to move together with measurable progress.

Concept diagram

CONVERGENCE DECISION FLOW

property bucket -> engine strategy -> helper invariants -> convergence audit -> closure

Metric graph

CONVERGENCE BURNDOWN

open hard properties    ███████
inconclusive aging      █████
closed with audit       ████████

Metrics and artifacts to collect

• engine effectiveness by property class

• induction and helper-lemma success ratio

• stalled-property aging dashboard

• runtime vs closure-quality movement

Mini case study

A stalled set closed only after case-splitting by mode and auditing fairness assumptions for realism.

Debug branches

• Bucket properties by structure and intent before tuning.

• Inspect proof core stability, not runtime alone.

• Reject speed gains that reduce legal reachability.

Senior review question

Ask: which requirement intent is proven, under which assumptions, and what residual risk remains?

Key takeaways

• Tie each proof claim to assumption boundaries and reachability evidence.

• Prefer minimal reversible fixes and preserve legal behavior visibility.

Common pitfalls

• Treating runtime reduction as proof-quality improvement without audits.

• Declaring closure while critical covers remain unreachable.

• Using broad waivers instead of first-divergence root-cause ownership.