What this chapter covers

STA concepts turn the idea of “the chip must meet timing” into a path-based calculation. A timing path has a startpoint and an endpoint; data arrival is summed along the launch clock and data network, required time is formed from the capture clock and library constraints, and slack is required time minus arrival for setup (and the reverse sense for hold).

This chapter is the vocabulary layer used in every later STA discussion: path types (reg-to-reg, in-to-reg, reg-to-out, in-to-out), clock latency versus skew versus uncertainty, and why PVT corners change cell delay even when the netlist does not. Interviewers listen for whether you can name those pieces without mixing them up. Once this vocabulary is solid, delay calculation, Liberty tables, and PrimeTime reports stop sounding like unrelated jargon and start sounding like one arithmetic story.

Key takeaways

  • Arrival and required times are clock-aware numbers, not just combinational delay.
  • Positive slack means the check passes; negative slack is a violation that must be scoped to a path, view, and constraint.
  • Hold is not a slower version of setup — it uses different edges, often different libraries, and is independent of period in the usual single-cycle sense.
  • Clock uncertainty, latency, and skew are related but not interchangeable; using one word for all three is an interview red flag.

Interview angle: walk a register-to-register setup check out loud: launch edge, clock-to-Q, combinational and net delay, setup constraint, capture edge, then slack. If you cannot point to each term, you are not ready for PrimeTime report questions.

Open the PDF reader for path diagrams, slack equations, and PrimeTime-oriented sketches that make the arithmetic visual.

Previous: Introduction to STA · Next: Standard Cell Library · Related questions: What is slack?, What is a setup check?, What is clock skew?

Sample interview answers

What are startpoints and endpoints? A startpoint is where a path’s timing begins (clock pin of a launching flop, or an input port). An endpoint is where it is checked (capturing data pin, or an output port).

How is slack computed? Slack is required time minus arrival time for setup (max paths), and arrival minus required for hold (min paths). Positive slack means the check is met.

Why do setup and hold use opposite worst-case clock skew? Setup is worst when capture arrives early relative to launch. Hold is worst when capture arrives late. The same physical skew therefore helps one check and hurts the other.

Related questions: What is a timing path?, Clock skew, STA handbook

💡 Key Technical Topics Covered

  • Timing Path Types (Reg-to-Reg, In-to-Reg, Reg-to-Out, In-to-Out)
  • Data Required Time vs Data Arrival Time
  • Setup Slack and Hold Slack Equations
  • Process, Voltage, and Temperature (PVT) Corners

🎯 VLSI Physical Design Interview Questions Addressed

  • How do you calculate Setup Slack in a Flip-Flop to Flip-Flop path?
  • Why is hold timing independent of clock frequency?
  • What is the impact of temperature inversion on lower technology nodes?