open-source-silicon.dev

6_final.gds

image.png

6_final.odb

6_final_drc.rpt

Hi, I was running the OpenFASoC <https://github.com/idea-fasoc/OpenFASOC/tree/main/openfasoc/generators/ldo-gen|LDO Generator> flow but got some DRC errors in the final layout.

It seems some DFF cells are too close (are they overlapping?) in the attached image. The GDS and ODB for the same are also attached along with the DRC report. The technology is `sky130hvl` and routing layers are disabled for better visibility.
I also tried using the `CELL_PAD_IN_SITES_GLOBAL_PLACEMENT` variable of ORFS to increase the padding between cells but it did not work. Am I missing something or is there a new way to do this? Any help would be appreciated, thanks!

cc <@U0169AQ41L6>

Are the cells reported as overlapping in the detailed placement report? If not it could be a LEF issue where the size of the LEF doesn't match the GDS.

As a quick workaround you could actually try setting `CELL_PAD_IN_SITES_DETAILED_PLACEMENT`  (i think, check the OpenROAD Flow docs). The detailed placer is the tool that is responsible for making sure cells don't overlap like that

I'll check the dpl report and tell soon, but I forgot to mention that I did set both the global and detailed placement padding options.

3_5_opendp.json

3_5_opendp.log

I don't know what to look for exactly but I couldn't find anything regarding overlap

The gds and lef match in size but I did notice that the actual measurements I made for the lef in klayout (including nwell) are slightly higher than the size mentioned in the lef text.

```MACRO sky130_fd_sc_hvl__dfxtp_1
  CLASS CORE ;
  FOREIGN sky130_fd_sc_hvl__dfxtp_1 ;
  ORIGIN  0.000000  0.000000 ;
  SIZE  12.00000 BY  4.070000 ;
  SYMMETRY X Y ;
  SITE unithv ;```

It seems to be slightly larger in one dimension even after excluding the nwell layer

Well I can tell you that OpenROAD doesn't see the cells as overlapping. Not sure what the deeper issue is