• Previous message: Jeff W. Sondeen: "Re: process scaling"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Dear magic hackers,

   I'm in the middle of recoding the CIF and GDS input and output sections
to new heights of sophistication.  This includes rescaling everything in a
design so that the grid space (magic units per lambda) can be changed at any
time, including during CIF and GDS input.  This means that when magic reads
non-magic-generated CIF and GDS, and finds geometry on finer scales than 1/2
lambda, it will redefine its own grid such that the grid is fine enough to
accept all input without altering any geometry.

   I also implemented a new DRC rule called "widespacing" to accept rules such   
as "if m1 width > 10um, then spacing to unconnected m1 must be at least 0.6um"
that seem to abound in modern processes.  It's actually a rather elegant
implementation, making use of the same code as the drc extensions' "maxwidth"
rule.  At least, I hope it will be regarded as elegant.

   Concerning the Darwin port emailed to the magic-dev group:  Please don't
make any other changes until I write back the changes mentioned above.  I'll
take care of the Darwin port additions when I'm finished.

							Regards,
							Tim

• Next message: Jeff W. Sondeen: "Re: -- SCN5M_SUBM.15.light.tech27"
• Previous message: Jeff W. Sondeen: "Re: process scaling"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]