https://open-source-silicon.dev logo
#osu
Title
# osu
t

Tuohang Zeng

03/13/2022, 2:41 AM
@User @User@User Hi, when I tried running DRC in klayout, it raised violations for the close nsdm shapes (red meshes) that are not merged. I found the nsdm layer to be 0.01um short compared to the site height (4.44um) . Also, do we have fill cells that have the nsdm and psdm layers? Thanks
t

Tim Edwards

03/13/2022, 3:25 PM
I can't tell anything from this. What are the cells shown here? FYI, all fill cells have NSDM and PSDM stretching the width of the cell.
t

Tuohang Zeng

03/13/2022, 5:42 PM
Those are the OSU 12T hs cells. From looking at the fill cells, I don't see NSDM and PSDM stretching across them. When I opened the inv_1 cell, as an example, the nsdm layer reached 0.01 um less than met1. It only reached 4.43um as opposed to 4.44um which is the site height. When the cells are placed on the layout in opposite orientations, they left a thin 0.02um wide space in the middle where the nsdm layers do not cover.
t

Tim Edwards

03/13/2022, 7:15 PM
Since you didn't say that you were using the OSU cells, I assumed that you were using SkyWater standard cells. If the OSU cells have issues with causing DRC errors when abutted in a certain way, then you need to take that up with James Stine, whose group designed the cells.
🙏 1
t

Teo Ene

03/13/2022, 9:27 PM
@User let me get back to you on Monday. Rushing to meet a deadline today.
There are two possible explanations. I can't remember off the top of my head how we handled PDSM in the end, because we went back-and-forth several times between manually painting it or using a rule.
@User just in case the error exists in some, but not all, cells, would you mind letting me know which cells are involved in that first screenshot?
t

Tuohang Zeng

03/13/2022, 10:55 PM
Hi @User, I can't speak for all the cells, but I think all the cells we encountered share this problem. A cell that has this issue is sky130_osu_sc_12T_hs__inv_1. Thanks!
m

mehdi

03/13/2022, 10:59 PM
@User we fixed this locally but it would be great when you have some cycles to update it on github. Also where should we file issues in the future?
t

Teo Ene

03/13/2022, 10:59 PM
Got it, thanks! I'll check and see what the issue is.
@User someone from the group should update it on github shortly, I'll ping everyone once I see what's wrong
🙌 1
As for where to file issues, let me go ask what the current location of our public repo is
It seems like this used to be the place?: https://github.com/google/skywater-pdk-libs-sky130_osu_sc.git But it's no longer up.
I'll ping the group; OSU needs some public-facing repo on GitHub for this.
t

Teo Ene

03/14/2022, 7:40 PM
That is correct. That is where the std cells (and their characterization) are located.
The issue is just that there is not a public-facing GitHub repo as well, where issues can be posted.
@User @User I've identified the issue (a mis-sized BOUNDARY layer, which we relied on to automatically draw NDSM/PDSM) and let OSU know about it.
Sorry, meant to post this yesterday, never managed to hit the "send" key.
t

Tuohang Zeng

03/15/2022, 4:05 PM
Thank you very much @User
2 Views