<@U017X0NM2E7> Hi David, Hi Tim <@U016EM8L91B> <@U...
# shuttle-precheck
s
@Mitch Bailey Hi David, Hi Tim @Tim Edwards @jeffdi I have disconnected all the gpio in connections and even then I am seeing OEB fails. attached the log files for the reference. Can you please guide to fix this issue. This issue was not found in the "initial_submit_04" run submitted re-used this database and fixed the FEOL error and re-submitted, now OEB errors are araised. Any update done in the submission or anything
m
@vks The reason the OEB check is failing is because the
vssa*
ports can not be found. It may be that they have been shorted to some other net. Looking at the
ext.log
, I can see a number of these lines.
Copy code
*.ext, line *: Number of subscripts doesn't match
I believe these occur when magic has problems extracting hierarchies (cross hierarchy devices or missing pins). We can try flattening these levels and see if it makes a difference. Can you add these lines to
EXTRACT_FLATGLOB
in your
lvs_config.json
file?
Copy code
*_CDNS_*
*Decoder_x*
*subtractor_x*
*bitcell_*
(don’t forget the `"`s and `,`s.
s
@Mitch Bailey I have shorted vssa1 and vssa2 intentionally.can we bypass this error?
t
@samarth jain: Avoid directly shorting two pins; you can add a metal resistor between one of the pins and the rest of the net to keep the nets connected but keep both the pin names in the netlist. You can also try
ext2spice short resistor
when extracting from magic; however, in that case magic will have to decide where to split the net and may not end up with the result you want.
s
Hi @Tim Edwards ,I have cadence not magic
t
You can still use the metal resistor.
m
@samarth jain Have you shorted vssa1 and vssa2 to something else? As long as one exists in the final netlist, the oeb check should work. You can check the netlist at
precheck_results/<tag>/tmp/ext/user*wrapper.gds.spice
.