Investigate why sky130hd/aes have large DRT runtime with smaller reduction values on GRT

[eder-matheus]

Subject

[Stage]: Global Router.

Describe the bug

The sky130hd platform uses a default layer resources reduction of 20%. For designs such aes, this reduction percentage seems too small, causing DRT to run many more iterations to complete the routing when compared to using a larger reduction.

In other words, the aes design benefits from a larger reduction value because it reduces the detailed router's number of iterations. We are not doing a good job in routability during GRT and need more reduction to compensate.

Investigate why the global routing result with 20% reduction produces so many more iterations in DRT.

Expected Behavior

GRT should produce a global routing result with good routability for DRT even when using smaller reduction percentages.

Environment

Not applied.

To Reproduce

In ORFS, modify the file flow/designs/sky130hd/aes/fastroute.tcl in its first line to apply 0.2 of reduction, instead of 0.4.

Relevant log output

No response

Screenshots

No response

Additional Context

No response