CTS

[EffeErreG4]

Is it possible to change the constrain on the maximum slew and skew for the CTS synthesis?

[maliberty]

@eder-matheus please comment.

[eder-matheus]

@EffeErreG4 CTS has the configure_cts_characterization command with a few options that may fit what you need. You can check the description for the command here: https://github.com/The-OpenROAD-Project/OpenROAD/tree/master/src/cts#configure-cts-characterization.

[EffeErreG4]

I already tried to explore this command for what concerns the maximum slew using the command "configure_cts_characterization -max_slew 2.0e-10" right before the "clock_tree_synthesis" one, but in the final log I always found a violation for a signal with maximum slew constrain of 1.51, which I assume is ns as expressed in other logs, which is the same result as if I wasn't using the command at all. Am I doing something wrong?
For what concerns the skew, is there a way to allow a more relaxed constraint and explore even more the functionalities of the program? Thanks for your help and availability.

[eder-matheus]

Could you provide a testcase with the script you've tested?

[EffeErreG4]

The testcase is the "jpeg_encoder" built in the sky130hd technology found here https://github.com/The-OpenROAD-Project/OpenROAD-flow-scripts/tree/master/flow/designs/sky130hd/jpeg
The "cts.tcl" file is mostly the same found in the OpenROAD-flow-scripts with the following changes:

configure_cts_characterization -max_slew 2.0e-10
clock_tree_synthesis -root_buf [list "sky130_fd_sc_hd__buf_16"]
-buf_list [list "sky130_fd_sc_hd__buf_16" "sky130_fd_sc_hd__buf_8" "sky130_fd_sc_hd__buf_4" "sky130_fd_sc_hd__buf_2" "sky130_fd_sc_hd__buf_1"]
-sink_clustering_enable
-sink_clustering_size 40
-sink_clustering_max_diameter 100000 \

The diameter is deliberately huge, to avoid this bottleneck in the clustering operation.
Unfortunately, attaching the file here, GitHub gives me a general error, if you want other files and you feel like to, you can write me at this email: s276848@studenti.polito.it