The python3 21_reram_simulation_test.py -t sky130 -v --num_banks=1 --num_cols=16 --word_size=16 --num_words=16

[Samarthjainabout]

The script gets stuck for 3 hours

[lekez2005]

I'm unable to reproduce this locally. Can you run with -vv instead for a more verbose log? i.e. python3 21_reram_simulation_test.py -t sky130 -vv --num_banks=1 --num_cols=16 --word_size=16 --num_words=16. You can attach the log file to a comment for review

[Samarthjainabout]

The error is weird. says the file is not there but when I check the file using less I can see it in the same path.I have attached the whole log folder as well.

reram_16_c_16_w16.zip

[lekez2005]

From those logs I can see that the temporary directory wasn't being adequately set when the SCRATCH environment variable is not defined. I have pushed a fix to use /tmp if $SCRATCH is not set. Can you test if that works?

[Samarthjainabout]

after updates still following errors.cant see error in log but warnings openram.log

[lekez2005]

Alright, looks like that fixed the previous issue. I'll need to take a closer look at either the magic version or the PDK version. I see you're using magic version 8.3.358 based on the logs while I'm still on 8.3.309. I'll get back to you once I test with version 8.3.358

[lekez2005]

I have been able to reproduce this bug using magic version 8.3.358. I don't really know what changed to break the extraction scripts. You can use magic version 8.3.309 until I further investigate how to correctly use the latest version of magic.

-- Edited link

[Samarthjainabout]

its 8.3.309 or 8.3.310?

I have been able to reproduce this bug using magic version 8.3.358. I don't really know what changed to break the extraction scripts. You can use magic version 8.3.309 until I further investigate how to correctly use the latest version of magic.

[lekez2005]

Yeah, sorry 8.3.309. Edited to reflect the correct link

[Samarthjainabout]

after changing magic,i get the following error.

openram.log

[lekez2005]

Did you modify the sense amp? This step was successfully run in the previous log file you posted which suggests to me that you've modified the sense amplifier

[Samarthjainabout]

I havent chnaged the sense

Did you modify the sense amp? This step was successfully run in the previous log file you posted which suggests to me that you've modified the sense amplifier

[lekez2005]

I havent chnaged the sense

Did you modify the sense amp? This step was successfully run in the previous log file you posted which suggests to me that you've modified the sense amplifier

I suspect the reason this didn't work is because you didn't specify to re-run extractions. By default, extraction is only run if a pex.sp file doesn't already exist since extraction takes a long time. The --run_pex explicit flag exists to enforce re-running extraction. FYI, if you make changes, --run_drc and --run_lvs should also be specified to ensure drc and lvs are run on the newest versions. The full command would look like: python3 21_reram_simulation_test.py -t sky130 -v --num_banks=1 --num_cols=16 --word_size=16 --num_words=16 --run_drc --run_lvs --run_pex

[lekez2005]

I have been able to reproduce this bug using magic version 8.3.358. I don't really know what changed to break the extraction scripts. You can use magic version 8.3.309 until I further investigate how to correctly use the latest version of magic.

-- Edited link

I found the issue: magic changed the extraction netlist's hierarchy separator in later versions from '/' to '.'. I have pushed changes to fix this so you should be able to use the latest version of magic again

[Samarthjainabout]

python3 21_reram_simulation_test.py -t sky130 -v --num_banks=1 --num_cols=16 --word_size=16 --num_words=16 --run_drc --run_lvs --run_pex

after I run this command, it gives same error? openram.log

[lekez2005]

python3 21_reram_simulation_test.py -t sky130 -v --num_banks=1 --num_cols=16 --word_size=16 --num_words=16 --run_drc --run_lvs --run_pex

after I run this command, it gives same error? openram.log

Just realized this is a different error from the one fixed by my magic-related changes. I am unable to reproduce this error from a fresh checkout of the repository. Can you try to reproduce this with a completely clean checkout from the repo? If it still happens, you can also attach the temp.sp spice file in the temp directory?

[Samarthjainabout]

Hi, I tried with latest git file and magic version 8.3.358 and got the same error. PFA Logfiles and temp.sp openram.zip

[lekez2005]

As discussed offline, this issue was because OPENRAM_HOME environment variable was different from the location of the actual code.