Description
We seem to be inserting too many TIE cells with repair_tie_fanout
Message [INFO RSZ-0042] exposes the amount of TIEs inserted at stage 2-1 from ORFS. Stage 2-1 is executed with script flow/scripts/floorplan.tcl with the command repair_tie_fanout.
/logs$ grep -rin "\[INFO RSZ-0042\]"
asap7/cva6/base/2_1_floorplan.log:51:[INFO RSZ-0042] Inserted 1021 tie TIELOx1_ASAP7_75t_R instances.
asap7/cva6/base/2_1_floorplan.log:53:[INFO RSZ-0042] Inserted 6822 tie TIEHIx1_ASAP7_75t_R instances.
I noticed we insert 75k TIEs on a private design which increases the design area by ~10%. Although I am not sure how much of the 10% is responsible to the TIE insertion alone.
Here are some plots showing the utilization progression, notice the increase on some designs for stage 2-1. This effect does not seem so evident on public designs.
This relates to issue https://github.com/The-OpenROAD-Project-private/OpenROAD-flow-scripts/issues/1565 which include plots for private designs.
Suggested Solution
No response
Additional Context
No response
Description
We seem to be inserting too many TIE cells with
repair_tie_fanoutMessage
[INFO RSZ-0042]exposes the amount of TIEs inserted at stage 2-1 from ORFS. Stage 2-1 is executed with scriptflow/scripts/floorplan.tclwith the commandrepair_tie_fanout.I noticed we insert 75k TIEs on a private design which increases the design area by ~10%. Although I am not sure how much of the 10% is responsible to the TIE insertion alone.
Here are some plots showing the utilization progression, notice the increase on some designs for stage 2-1. This effect does not seem so evident on public designs.
This relates to issue https://github.com/The-OpenROAD-Project-private/OpenROAD-flow-scripts/issues/1565 which include plots for private designs.
Suggested Solution
No response
Additional Context
No response