The Tcl reporting command report_phys_opt
provides details of each
optimization performed by phys_opt_design
at a very fine level of
detail. It must be run in the same Vivado session as
phys_opt_design
while the optimization history resides in
memory.
Therefore, if a report is desired, it is recommended to include the
report_phys_opt
command in Tcl scripts immediately following
the last phys_opt_design
command.
The reports are available only for post-placement phys_opt_design
optimizations. The reports are cumulative, reflecting all
phys_opt_design
optimizations, including those from multiple
runs of phys_opt_design
.
The following report example shows the first entry of a fanout optimization involving a register named pipeline_en. The following details are shown in the report:
- The original driver pipeline_en drives 816 loads and the paths containing this high fanout net fail timing with WNS of -1.057 ns.
- The driver pipeline_en was replicated to create one new cell, pipeline_en_replica.
- The 816 loads were split between pipeline_en_replica, which takes 386 loads, and the original driver pipeline_en, which takes the remaining 430 loads.
- After replication and placement of pipeline_en_replica, the WNS of pipeline_en_replica paths is +0.464 ns, and the WNS of pipeline_en paths is reduced to zero.
- The placement of the original driver pipeline_en was changed to improve WNS based on the locations of its reduced set of loads.