Could someone quickly check if OpenLane `flow tcl design spm open-source-silicon.dev #general

Could someone quickly check if OpenLane `flow.tcl ...

Manuel M

11/05/2022, 10:44 AM

Could someone quickly check if OpenLane

flow.tcl -design spm

with config

"MAGIC_EXT_USE_GDS": true

fails at LVS with

unmatched pins = 344

recently? After updating to the

tag=next

iic-osic-tools I see internal wires being interpreted as I/O ports in the exported

.spice

, which did not happen in the image a month ago. open_pdks 0059588, OpenLane a0b19fa, Magic 8.3.331

Manuel M

11/05/2022, 10:51 AM

Trying to find out if this happens in a different environment while I can't easily jump between commits inside of the docker image.

Copy code

.subckt spm VGND VPWR _000_ _001_ _002_ _003_ _004_ _005_ _006_ _007_ _008_ _009_
+ _010_ _011_ _012_ _013_ _014_ _015_ _016_ _017_ _018_ _019_ _020_ _021_ _022_ _023_
+ _024_ _025_ _026_ _027_ _028_ _029_ _030_ _031_ _032_ _033_ _034_ _035_ _036_ _037_

Vijayan Krishnan

11/05/2022, 10:53 AM

MAGIC_EXT_USE_GDS

if enabled facing LVS issue with latest OpenLane. If disabled there is no LVS issue reported.

Manuel M

11/05/2022, 10:58 AM

Yes, export with

DEF/LEF

instead of

GDS

is working, will use it as workaround. Thanks @Vijayan Krishnan

Harald Pretl

11/05/2022, 11:00 AM

@donn @Tim Edwards @Vijayan Krishnan Do you plan to restore the earlier behaviour so that the

MAGIC_EXT_USE_GDS

switch works again as intended?

Vijayan Krishnan

11/05/2022, 11:20 AM

Please report as github issue, so it will be handled and fixed in OpenLane master

✅ 1

Open in Slack

Previous Next