The OverCon in Encounter,urgent!!!

[david_zheng]

Dear all,

Please check my design congestion analysis as below,the overcon of ME2 looks very serious,and after global/detail route,the number of DRC vios greater than 5K,
the quality of routing is very bad.my question is


1>why there are so many OverCong in ME2 ? while OverCon of some other metal are very low,how to control OverCong for a certain metal layer in design?
2>Any other reason will impact the OverCong,floorplan? or others? how can solve it respectively?

Thanks and regards,
david

 

[rca]

do you have 5 or 7 metal avaliable?
Metal2 is used to power the std cell?

what's your std cell density?

 

[david_zheng]

do you have 5 or 7 metal avaliable?
Metal2 is used to power the std cell?

what's your std cell density?

My process is 1p7m,but I used ME6/7 only for power ring/strip,and my placement density is 0.526,please advice,thanks.

 

[rca]

you don't believe the tool will use the top layer metal to fix antenna? or do you prefer to insert diode.

do you used the correct techno file provide by the std cell vendor? some time the std cell vendor modify a bit the track alignment to fit with their std cell factor, and optimize the track routing avaliable.

 

[david_zheng]

you don't believe the tool will use the top layer metal to fix antenna? or do you prefer to insert diode.

do you used the correct techno file provide by the std cell vendor? some time the std cell vendor modify a bit the track alignment to fit with their std cell factor, and optimize the track routing avaliable.

I have enabled M6/M7 for routing,but can't improve total overcong and would you please show me an example about what factor of std cell and tech file will effect PNR,thanks a lot.

 

[rca]

have you checks if the std cell vendor provides also a lef techno file ?

 

[cop02ia]

Hi,

I think it's quite hard to fix congestion only on certain metal layers just based on statistics. It's more intuitive to fix according to area e.g. placement density control, create routing channels.

My guess is that congestion on Metal 2 is due to vertical connection to standard cell pins.
Try buffering standard cells so they have some space between them horizontally (say 10% size of smallest inverter).

It's like creating halo on the left & right of a macro but instead, you do it on standard cells (check with tool vendor how to do this). I'm assuming that your power/ground rails on Metal 1 for standard cells are going horizontally.

Best regards.

 

[friend_333]

hey

what is the intial uilization ?