Logo
VLSI Physical Design  ›  Ch 9. Crosstalk & Signal Integrity

Clock Frequency and Signal Integrity

Yes. Changing the clock frequency changes the crosstalk arrival windows and the resulting crosstalk noise bumps when a steady signal passes near a clock edge. As a result the delay of crosstalk-affected cells changes and more setup and hold violations appear, which the tool will work to fix if there is still optimization room at the route stage.

KEY Yes - a shorter period shifts crosstalk arrival windows, changing affected-cell delays and surfacing more SI-driven violations.

Preventing Signal Integrity Issues

Placement:

  • Reduce congestion or use SI-aware placement.
  • Avoid high cell-density regions.
  • Use place_opt with congestion and area-recovery options.
  • Avoid low-drive-strength cells, since they tend to become victim nets.
  • Keep very high drive-strength cells on a dont-use list, since they tend to become aggressor nets.
  • Control the maximum transition constraint on the design to help prevent crosstalk; it is technology- and library-dependent, so find the best trade-off between a low max-transition limit and congestion, and relax it during post-route optimization if needed.
  • Use the maximum net length constraint in ICC to prevent very long wires and reduce crosstalk.

CTS:

  • Apply NDR rules to the clock network so it is less sensitive to crosstalk.
  • Apply extra spacing between the clock network and signal nets.
  • Because clock nets are high-frequency and act as strong aggressors, shield them with ground wires.
  • Avoid placing clock gaters very close together by adding padding, since they act as aggressors to adjacent signal nets.

Routing:

  • Use SI-aware (crosstalk-aware) detailed routing.
  • The router can prevent crosstalk during global routing and track