Generating SRAM gets stuck in extraction process of magic

[FriedrichWu]

Describe the bug
Hello,

I'm trying to generating 2kByte Sram, but the process gets stuck in run_ext for nearly one day. This issue seems to occur more when sram size is large. I could make the whole sram generation process be able to finish by disabling lvs & commenting the stimulus and measurement part, but this is of course not a very good solution.

This is a similar issue like here->#245

And in the .ext.err file I find a lot of warnings, besides "unknown layers" and "boundary was redefined", there is also error shows "Error: Asymmetric device with multiple terminals!", which is also mentioned here->https://web.open-source-silicon.dev/t/424007/i-m-trying-the-sky130-support-in-openram-using-the-example-c

Version
The current latest version

Expected behavior
The process could finish without error

Logs
myconfig_sky.ext.err.zip

[mguthaus]

This sounds like a run time bug with magic. There isn't much we can do to speed up the extraction on the OpenRAM side.

My only questions are how much memory you have and are using. Especially if you are running it in Docker or a VM.

[mguthaus]

The asymmetric device message is expected since the SRAM cells have an asymmetric transistor or two.

[FriedrichWu]

This sounds like a run time bug with magic. There isn't much we can do to speed up the extraction on the OpenRAM side.

My only questions are how much memory you have and are using. Especially if you are running it in Docker or a VM.

Thanks for your quick reply. I'm running on Linux sever, it has around 124 GB memory available and another 37 GB for swap