• Previous message: David Fletcher: "Re: Cadence interface to Magic?"
• In reply to: Rajit Manohar: "Re: Cadence interface to Magic?"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

didn't see any followup to this, so i'll "stab" it.

the magic 'router' is a very simple thing whose intent is really for
perhaps getting signals from your core to your pads, but even then it's
usually faster to route them manually than prepare a netlist file for
the router.

it requires a simple grid to handle all layers, so like 8x8 lambda if
your (symbolic) contacts (like m2c) are 4x4 and metal spacing is 4
lambda, for example.  then i remember strange behavior unless all pins
had their LL (lower-left) corner on the grid.

i can dust off an example if you're still interested but i don't think
the stock magic router will do anyting satisfactory with true standard
cell routing. also, there's no autoplacer in magic.

/jeff


Rajit Manohar writes:
 > 
 > Hi,
 > 
 > Magic can produce GDS stream output and read GDS stream input (using
 > the :calma command) as long as the technology file contains sections
 > for calma I/O.
 > 
 > Magic has some rudimentary routing capability. I normally use the
 > interactive router (:iroute) to route individual nets, but normally I
 > don't use it for large autorouting problems. Maybe others have used it 
 > for this purpose?
 > 
 > -Rajit

• Next message: Ronny Bockstaele: "question on the technology file"
• Previous message: David Fletcher: "Re: Cadence interface to Magic?"
• In reply to: Rajit Manohar: "Re: Cadence interface to Magic?"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]