<@U016Q3JTPA7>: Sorry, I've been ignoring this cha...
# sky130
t
@User: Sorry, I've been ignoring this channel most of the day in order to get actual work done. But the slightly longer answer is: Qflow works fine but it does suffer from older algorithms for placement and routing that tend to max out at about 50k gates or so. So OpenROAD is the future and I've been a vocal proponent of it since its inception. Qflow also does not work well with high-density standard cell sets, and there are a few other problems with the sky130 process that I have not yet looked into that are interfering with routing. We had the same sort of issues with openlane/OpenROAD; it's just that I haven't had the time to look into them. I would guess that eventually I can get qflow working fine for small projects and maybe continue development of the flow, especially to try to integrate the OpenROAD tools into it. Right now I can get it to run synthesis and placement in sky130, but routing tends to stop with a lot of route failures. I've been using qflow a lot to understand the sky130 standard cells better.
👍 1
k
I am trying a simple design using sky130 on Open Galaxy but when I run it through qflow, it is giving me route fail. I saw this post and was wondering if the issue was resolved? If yes, could you provide me with the solution?
t
The qrouter router in qflow is not well suited to the geometry of the sky130 standard cells, and I basically gave up trying to do anything about it because OpenROAD/openlane was available. There is, I think, a good possibility that qflow will work well with the OSU standard cell sets for sky130, but I haven't spent any time looking into that.
k
Thank you for your advice. I got the flow running for my design using Openlane. I was trying to verify it but I got an error (see attachment). Is there any documentation on verifying the layout?