I am currently developing a PDK for a foundry that is limited by layers. It is not a silicon technology; it is based on organics. I have only two metal layers, which are both used inside the transistor designs as well. Therefore, I have the following questions: 1- Do I need tie cells in my library? 2- Is there any option to get vertical padding? Since I have limited metals that have been already used inside the transistor structure, I need to move the routing outside the designs. 3- Is there any straightforward way to adapt the PDK to OpenLane? The documentation is very complicated.

1. Yes. We can adjust that, however, if your process does not require tie cells for some reason. 2. I don't entirely understand what you mean by this. Do you mean an obstruction on top of certain cells? I believe the router will already dodge any metals that are inherently inside cells/transistors… 3. Not really. The best way continues to be to start with the sky130 configuration file as a skeleton and just replace the files as you go.

Is there any command that stop openlane to rotate cells and only mirror them by Y axis?

@Ashkan With regards to "Is there any command that stop openlane to rotate cells and only mirror them by Y axis?" - OpenLane gets this information from cell LEF file, specifically the symmetry field:

MACRO sky130_ef_sc_hd__decap_12
  …
  SYMMETRY X Y R90 ;

If a cell can only be mirrored across the Y axis,

SYMMETRY Y;

might be what you're looking for. Obviously, I have no idea how you're generating your LEFs, but there should be an option to do that. --- There is currently no option afaict to add padding between rows in OpenROAD, which OpenLane relies on for floorplanning and PnR. A dirty solution would be a custom step that fills every other row with blank cells, effectively leaving a whole row of padding between cells. @Matt Liberty Is there such an option in OpenROAD (and if not, can it be added?)

@donn In our LEF file the cell symmetry has been defined like:

SYMMETRY Y ;

But it still rotating.

Mmyeah we'd need the files to debug

This looks to be a duplicate of an already reported OR issue

Then are you going to fix this? @Matt Liberty

I already suggested a workaround. It is not a high priority as it is not a normal usage

Sorry, what's the workaround? (I can't find the issue 😅)

Simply delete the rows you don't want to use

Reasonable. Thanks Matt