Explain the interaction between POCV and clock reconvergence pessimism removal on a clock path - do they double-count?
From PDVerse PrimeTime STA Interview Guide · pdVerse Mentor Guide
Ten-second Interview Answer
No - they correct different artifacts. CRPR removes the double-count on the clock segment shared by launch and capture; POCV graph-merging removal removes the stochastic-max inflation at converging gates. Both must be active on a clock path.
Complete Technical Explanation
They address different pessimisms and are complementary rather than redundant. CRPR removes the shared-clock-segment double count that OCV and POCV introduce when the launch and capture clock paths overlap, since the analysis would otherwise charge the common segment two different delays at once. POCV graph merging pessimism removal removes the stochastic-max inflation at converging gates, where the statistical max of several fan-in distributions exceeds any single delay. On a clock path you can genuinely have both: the common clock portion needs CRPR to reconcile launch against capture, and any reconvergent clock structure, or converging data at the endpoint, needs graph-merging-pessimism removal. They cannot double-count because the mechanisms they correct are different - CRPR is about how a shared segment is treated on the launch side versus the capture side, while graph merging is about the statistical operation of merging fan-in distributions. There is one subtlety worth having ready: under POCV, the delays CRPR reconciles are themselves distributions with a mean and a sigma, so CRPR operates statistically too. The common path pessimism is computed on the distribution, and the clock reconvergence pessimism line coexists with the statistical graph pessimism line in a POCV report_timing -variation . An accurate POCV clock path needs both corrections enabled and understood; a candidate who conflates them - thinking graph-merging removal subsumes CRPR, or the reverse - misses that they fix orthogonal artifacts.
In Pd / Signoff
On a POCV clock path debug you open report_timing -variation and look for both lines: the clock reconvergence pessimism entry and the statistical graph pessimism adjustment. One missing line tells you which correction is disabled or not applying, and that is usually why a deep clock path looks unaccountably pessimistic.
Common Trap
Assuming that because POCV already removes merging pessimism, CRPR is redundant and can be relaxed - or the reverse, that CRPR covers the convergence artifacts so the statistical adjustment can be ignored. Disabling either leaves a real artifact in place and inflates clock path pessimism.
Expect Next
Under POCV, what exactly is the common path pessimism computed on when the shared clock segment's delay is a distribution?
Topic And Primetime Commands
Variation / POCV Distributions. report_timing
Keep building interview depth
Continue practising STA
Continue learning free
Get a practical low-power chapter
Receive the existing “Low Power and Multivoltage Fundamentals” PDF chapter and its download link by email.
Continue practising