English
Menu

Timing analysis and CDC

When a setup multicycle path of N (N ≥ 2) is set, a hold multicycle path of N − 1 is normally set as well. Why?

  1. ABecause the hold check edge also shifts later along with setup and becomes far too strict, so it must be moved back to its original position
  2. BTo shift the hold check edge even later and increase hold margin
  3. CTo exclude the path's hold check from the analysis
Show answer and explanation

Answer: A. Because the hold check edge also shifts later along with setup and becomes far too strict, so it must be moved back to its original position

By default, STA checks hold at the edge one before the setup capture edge. Making setup two cycles moves the hold check edge from time 0 to one period later, so with a 2.0 ns period the data would violate hold unless it arrived 2.0 ns or more late. Specifying a hold multicycle path of N − 1 = 1 moves the check edge back to time 0, giving the same conditions as an ordinary path. It does not remove the hold check.

Practice this topic Back to topic

Books on Timing analysis and CDC Ad

As an Amazon Associate I earn from qualifying purchases.

More Timing analysis and CDC questions

  1. For which path is it appropriate to set a false path (set_false_path)?
  2. With a 2.0 ns clock at 50% duty cycle, a path launches on the rising edge and captures on the falling edge. With a setup time of 0.1 ns and zero skew, what is the maximum allowed sum of the clock-to-output delay and combinational logic delay?
  3. With a 2.0 ns clock, a path runs from a flip-flop that launches data on the rising edge at time 0 to a latch that opens (becomes transparent) at 1.0 ns and closes at 2.0 ns. If the data arrives at the latch at 1.4 ns, how much time has been borrowed at this latch (time borrowing)?
  4. In a circuit that stops the clock with a two-input AND gate of the clock and an enable signal, what does the STA clock gating check verify?
  5. With a 2.0 ns clock, set_input_delay 0.6 ns is set on an input port. If the internal flip-flop's setup time is 0.1 ns and clock latency and skew are ignored, what is the maximum delay allowed for the on-chip path from the input port to the flip-flop?