When a timing path shows a large negative slack, what are the two canonical fixes, and how do you choose between them?

Short Answer

For a violating R2R path, the two canonical fixes are: swap to faster cells along the path (if faster library variants exist), or split the path by inserting a register partway through, creating a new, shorter R2R endpoint. Either fix requires re-running LEC between the modified netlist and the golden RTL to confirm functional equivalence -- this same swap-vs-split methodology is the core approach used in post-PD ECO timing closure generally, not just this one worked example.

Technical Explanation

• Fix 1: swap to faster cells -- replace cells on the violating path with faster library variants if available. Simple, but limited by what the library actually offers.
• Fix 2: split the path (register insertion) -- register the output partway through the path, creating a new, shorter R2R endpoint out of what was one long path.
• Splitting genuinely changes the pipeline structure (adds a cycle of latency at that point), while swapping cells does not -- this is a real architectural tradeoff, not just a timing-closure detail.
• Either fix requires re-running LEC (logic equivalence checking) between the modified netlist and the golden RTL to confirm functional equivalence -- a timing fix that silently breaks function is not actually a fix.

What To Check

• Warning sign: a reported R2R path has large negative slack (well beyond what cell swapping alone is likely to recover).
• Inspect: check whether faster library variants for the path's cells actually exist and would recover enough slack, or whether the gap is too large for sizing alone.
• Correct: if sizing can't close the gap, consider path splitting -- but confirm the added pipeline latency is architecturally acceptable before committing to it, and re-run LEC after either fix.

Healthy, Suspicious & Hard-stop Results

• Expected: the chosen fix (swap or split) recovers positive slack, and LEC confirms the modified netlist remains functionally equivalent to the golden RTL.
• Investigate: cell swapping alone doesn't close the gap even with the fastest available library variant -- this is the signal to consider path splitting instead of continuing to search for a bigger swap.
• Stop: a timing fix (either type) is applied without a subsequent LEC re-run -- a timing-passing but functionally-broken netlist is not a real fix.

What The Interviewer Is Testing

Whether you know both canonical fixes and the real tradeoff between them (added latency vs. limited by library speed), and that LEC re-verification is mandatory after either one.