I noticed that the v0.2 of openlane creates gds in magic and klayout subdirectories. What is the reason for the two versions?

designs/spm/runs/test/results/klayout/spm.gds
designs/spm/runs/test/results/magic/spm.gds

The binary files differ, but that might just be the timestamp. Are there other intentional differences between these files? I get a clean device level LVS run on the magic gds, but the klayout version has a pin mismatch - possibly due to multiple texts.

VPWR_uq1                  |VPWR **Mismatch**

I ran LVS using the -lvs option in flow.tcl

./flow.tcl -design spm -tag lvs -lvs -gds designs/spm/runs/test/results/magic/spm.gds -net designs/spm/runs/test/results/lvs/spm.lvs.powered.v
./flow.tcl -design spm -tag lvs2 -lvs -gds designs/spm/runs/test/results/klayout/spm.gds -net designs/spm/runs/test/results/lvs/spm.lvs.powered.v

#verification-be #openlane

@User: Firstly, magic doesn't deal very well with the klayout generated GDS-II. Secondly, the two GDS-IIs are indeed different, and you can check the difference with the XOR result stored in the results/klayout folder. Finally, the aim of adding it is to allow the community to revise the klayout tech and drc files present under open_pdks. There are some layers/purposes that differ. We're still experimenting with the DRC rules coverage of the existing klayout DRC rule set. The final goal is for the two GDS-IIs to be identical. Still, we are (secretly) relying on the community to hunt down the differences and PR fixes to both openlane and open_pdks, which became more accessible with the new development structure.