There are some lacking steps in IHP port to “OpenROAD -flow-scripts” flow where
filler rules for GatPoly layer are not specified. I have added a respective
section to the fill.json file to drive the filer however it generates fill shapes
without considering existing connections on this layer (see picture below).
Any idea how to address this issue ?
a
Austin Rovinski
04/12/2024, 5:03 PM
I'm guessing GatPoly is a substrate layer? In that case, there is no way to perform fill in OpenROAD. OpenROAD only does fill for metal layers. It is expected that substrate density rules are satisfied by filler cells, or alternatively an external tool like KLayout.
If you really want to make it work with OpenROAD, I could see a possibility of hacking the LEF file to include the GatPoly layer. The main problem is that OpenROAD has no visibility into any layers which aren't present in the LEF file.
m
Matt Liberty
04/15/2024, 5:15 PM
Wouldn't this normally be covered by the filler cells?
k
Krzysztof Herman
04/23/2024, 5:38 PM
@Austin Rovinski seems that OR fills booth GatPoly and Active layers independent the USE_FILL macro, which triggers metal fill only. The question is if we can somehow control de density of this fillers. And yes booth belong to the substrate layers (FEOL). Any guess ?
Krzysztof Herman
04/23/2024, 5:41 PM
And both fills: GatPoly and Active are generated without any shorts (sems to be correct)
m
Matt Liberty
04/23/2024, 5:44 PM
The fill cells are required for well continuity and there is no option to control their density.
Linen is a search-engine friendly community platform. We offer integrations with existing Slack/Discord communities and make those conversations Google-searchable.