@Tim Edwards When I tried to extract a spice file from a hierarchical circuit (integrator_circuit.png), I encountered an error stating, 'cannot read extract file sky130_fd_pr__pfet_01v8_M479BK.ext'. Do you have any suggestions for solving this problem? By the way, I did not encounter the error when performing extraction on a circuit with two inverters (two_inverters.pdf). Also, I have checked that LVS was OK for this two inverter circuit. Thanks in advance.

@Yusuke Sakemi do you get the same results if you close magic, reopen the cell and re-extract?

@Mitch Bailey Yes, I just tried it, but I had the exactly same error. I atttached .ext file of the top cell.

@Yusuke Sakemi looks like the missing cell is in a subcell because it doesn’t show up in the top

ext

file. Can you

grep sky130_fd_pr__pfet_01v8_M479BK mag/*.mag

?

@Yusuke Sakemi: Can you post all the commands you used for the extraction?

@Mitch Bailey Thanks for your help. Using grep command, I found "sky130_fd_pr__pfet_01v8_M479BK" in ./passgate/passgate.ext. However, the device is not found in ./passgate/passgate.mag. Did this discrepancy cause the error? But in hierarchical design, I think the top cell .mag uses other .mag files, is it right? Thanks for your help. By using the grep command, I found 'sky130_fd_pr__pfet_01v8_M479BK' in './passgate/passgate.ext'. However, the same device is not found in './passgate/passgate.mag'. Could this discrepancy have caused the error? But in a hierarchical design, I believe the top cell '.mag' file uses other '.mag' files, right? @Tim Edwards I used the following commands to extract the spice netlist: extract do local extract all ext2spice lvs ext2spice

@Yusuke Sakemi > But in a hierarchical design, I believe the top cell ‘.mag’ file uses other ‘.mag’ files, right? Yes, that is correct. Is there a

sky130_fd_pr__pfet_01v8_M479BK

in any other file? Can you check the timestamps on the files? Is there an

ext

file that is much older than the corresponding

mag

file?

I wonder if there is just something caused by the directory "passgate" having the same name as the cell "passgate". Can you please tarball the layout for me? I would like to try to reproduce the result. If there is some directory/cell name confusion, then it should be easy to reproduce the error.

@Mitch Bailey @Tim Edwards I could find

sky130_fd_pr__pfet_01v8_M479BK

only in the following three files: passgate/passgate.ext, comp.out, passgate_magic.spice. I noticed that passgate.mag is newer than passgate.ext. So, I saved passgate.mag and extracted passgate.ext. Now, the error is gone, and the top cell LVS has also passed. Thank you for your support!

@Yusuke Sakemi Great! I would have though that

extract all

would force an extraction of all the subcells, but maybe I don’t understand that. @Tim Edwards does the expand region matter in extraction?

extract

will load all subcells and doesn't care if the design is visually expanded or not.