A path fails hold only after you enabled crosstalk analysis. Could CRPR be involved, and how would you investigate?
From PDVerse PrimeTime STA Interview Guide · pdVerse Mentor Guide
Ten-second Interview Answer
Possibly. If it is a zero-cycle hold check, crosstalk on the common clock segment should be CRPR-removed, and if that credit is missing you see false hold pessimism. The key discriminator is whether the check is zero-cycle.
Complete Technical Explanation
CRPR could be involved, but only under one condition. If the check is zero-cycle - the same edge launches and captures, as with a divide-by-2 or certain adjacent-flop structures - then crosstalk-induced delay on the common clock segment should be removed by CRPR, because the aggressor perturbs launch and capture identically. If that credit is not being applied correctly, the extra hold pessimism you are seeing is false. Investigate in order. First confirm the check really is zero-cycle by reading the launch and capture edges out of report_timing; do not infer it from the topology alone. Second, run report_crpr and see whether the crosstalk delta on the common segment is actually being removed. Third, verify timing_remove_clock_reconvergence_pessimism is true, since the whole mechanism is off otherwise. Fourth - and this is the branch people skip - if the check turns out not to be zero-cycle, then CRPR is behaving correctly by not removing the crosstalk delta, because launch and capture are different edges and the aggressor's effect on each is not guaranteed to match. In that case the hold failure is likely real signal integrity, and you should pursue SI fixes such as shielding or increased spacing rather than hunting for a missing pessimism credit. The zero-cycle question is the discriminator that tells you which of those two entirely different debugs you are in.
In Pd / Signoff
This is post-route hold debug once crosstalk analysis is switched on and the violation list grows. For each new failure you read the launch and capture edges, then either chase a missing CRPR credit or route the path to the SI team for shielding and spacing. Answering the zero-cycle question first is what keeps you from spending ECO budget on pessimism.
Common Trap
Assuming any crosstalk-induced hold failure on a shared clock segment must be CRPR pessimism. For non-zero-cycle checks CRPR correctly leaves the delta in, because different edges can genuinely see different aggressor activity, and removing it would be optimistic. The inverse error is shielding a zero-cycle path whose delta should have been credited back.
Expect Next
For a non-zero-cycle check, what would you have to prove about the aggressor before you could argue the delta is a double-count?
Topic And Primetime Commands
Operating Conditions / General. report_timing, report_crpr, timing_remove_clock_reconvergence_pessimism
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