route_design
# Unroute all the nets in u0/u1, and route the critical nets first
route_design -unroute [get_nets u0/u1/*]
route_design -delay -nets [get_nets $myCritNets]
route_design -preserve
The strategy in this example script illustrates one possible way to address timing failures due to congestion. In the example design, some critical nets represented by $myCritNets need routing resources in the same device region as the nets in instance u0/u1. The nets in u0/u1 are not as timing-critical, so they are unrouted to allow the critical nets
$myCritNets
to be routed first, with the smallest possible delay.
Then route_design -preserve
routes the entire design. The -preserve
switch preserves the routing of $myCritNets while the unrouted u0/u1 nets are
re-routed. Table 2-12 summarizes the commands in the example.