Dear @Harald Pretl I wanted to ask if you could update the MAGIC VLSI commit in the Docker file, as various issues have been solved in the latest version. It would help a lot 😄 Best regards, Vicente Osorio

The

magic

version is based on what OpenLane wants to see. When we release a new Docker version, we take the latest OpenLane and and thus the respective magic version. This means that magic is relative current, but not necessarily the latest one. Going to just the last commit would cause quite a few unwanted side effects, so we are cautios.

It is not too old, but a lot of things regarding parasitic extraction have been reworked and corrected since that commit in tag

latest

.

Alright, I see how current we can go. Which magic tag would you like to see (as a minimum)?

I think commit

73398e7e0b5fff7667ea3ef4e4687ce1ea5bdfd8

should be a minimum.

Not good. Latest OL (2023.03.08) has

1a3caee3761d3ad4d4d4a7e931137b21a608a92d

Thank you very much @Harald Pretl!!! 😄

Not sure whether this helps, but at least it is documented somewhere 😉

Dear @Harald Pretl, thank you for helping me by sending that issue, now OpenLane updated the MAGIC commit they use (OpenLane commit

dc8e6bb151df7ee2b93d7e9d6a15abf6f224db4d

) 😄. Is it possible to now update the commit tag in the Dockerfile? Best regards, Vicente Osorio Rivas

@Vicente Osorio Yes, I am working on the March release. However, currently OpenLane is broken, and I would need a fix for it. @donn (the KLayout write issue).

https://github.com/The-OpenROAD-Project/OpenLane/pull/1680 -- I'll fast-track it

@donn Merged and tested: Now we have consistency in that OL1 fails with the same error as OL2 😳 OL1:

[STEP 29]
[INFO]: Streaming out GDSII with KLayout (log: spm/runs/RUN_2023.03.16_11.57.50/logs/signoff/29-gdsii-klayout.log)...
[ERROR]: during executing: "python3 /foss/tools/openlane/2023.03/scripts/klayout/stream_out.py --output /foss/designs/spm/runs/RUN_2023.03.16_11.57.50/results/signoff/spm.klayout.gds --lyt /foss/pdks/sky130A/libs.tech/klayout/tech/sky130A.lyt --lym /foss/pdks/sky130A/libs.tech/klayout/tech/sky130A.map --lyp /foss/pdks/sky130A/libs.tech/klayout/tech/sky130A.lyp --top spm --with-gds-file /foss/pdks/sky130A/libs.ref/sky130_fd_sc_hd/gds/sky130_fd_sc_hd.gds --input-lef /foss/designs/spm/runs/RUN_2023.03.16_11.57.50/tmp/merged.nom.lef /foss/designs/spm/runs/RUN_2023.03.16_11.57.50/results/routing/spm.def |& tee /dev/null /foss/designs/spm/runs/RUN_2023.03.16_11.57.50/logs/signoff/29-gdsii-klayout.log"
[ERROR]: Exit code: 1
[ERROR]: Last 10 lines:
'NoneType' object has no attribute 'cell_index'
child process exited abnormally

OL2:

[INFO] GDS Write Complete
────────────────────────────────────────────────────────────────── GDS-II Stream Out (KLayout) ───────────────────────────────────────────────────────────────────
'NoneType' object has no attribute 'cell_index'
[12:00:44] ❌  The following error was encountered while running the flow: Command '['/usr/bin/python3',                                           __main__.py:159
           '/usr/local/lib/python3.10/dist-packages/openlane/scripts/klayout/stream_out.py',
           'spm/runs/RUN_2023-03-16_12-00-13/23-openroad-fillinsertion/spm.def', '--output',
           'spm/runs/RUN_2023-03-16_12-00-13/28-klayout-streamout/spm.klayout.gds', '--lyt',
           '/foss/pdks/sky130A/libs.tech/klayout/tech/sky130A.lyt', '--lyp', '/foss/pdks/sky130A/libs.tech/klayout/tech/sky130A.lyp', '--lym',
           '/foss/pdks/sky130A/libs.tech/klayout/tech/sky130A.map', '--top', 'spm', '--input-lef',
           '/foss/pdks/sky130A/libs.ref/sky130_fd_sc_hd/techlef/sky130_fd_sc_hd__nom.tlef', '--input-lef',
           '/foss/pdks/sky130A/libs.ref/sky130_fd_sc_hd/lef/sky130_fd_sc_hd.lef', '--input-lef',
           '/foss/pdks/sky130A/libs.ref/sky130_fd_sc_hd/lef/sky130_ef_sc_hd.lef', '--with-gds-file',
           '/foss/pdks/sky130A/libs.ref/sky130_fd_sc_hd/gds/sky130_fd_sc_hd.gds']' returned non-zero exit status 1.
           ❌  OpenLane will now quit.

god,

@donn KLayout is version 0.28.2

alright i'll look into it

No problem, I’ll do, but I have tested this yesterday, same fail, so I stepped back to 0.28.2. But I test again, and go to 0.28.5

0.28.5 is what we use in our (just released) containers, so yeah

@Vicente Osorio I have pushed a new release, tagged

latest

and

2023.03

. this should have a new magic version, plus a few other goodies (like openlane2).

Thank you very much @Harald Pretl and @donn!!!😄

Yw! Please give OpenLane 2 a shot as well, it's the nicest open secret we have so far. 🙂