Magic Mailing List |
|
From: Jeff W. Sondeen (sondeen AT rcf-fs DOT usc.edu) Date: Fri May 12 2000 - 17:51:56 EDT
R. Timothy Edwards writes: > Jeff, > The quick code that I set up for doing the half-lambda grid is > not well-developed, and I'm aware of all the problems you mentioned. > The way it was set up, the "scalegrid" command scales all magic > units by the specified ratio, and reflects the new scale in both the > cifoutput/cifinput and drc tech. Magic units are reported as grid, > not lambda, and are saved this way. Unless the grid-vs.-lambda is > specified somehow in every output magic file, there's no way to know you can do this with the first line of the .mag file; currently the "technology" is "scmos" but you could say, "scmos*1/2" or something -- then magic makes sure you don't mix "technologies". > what scale the magic file is supposed to be in. So there's one > thing that needs to be added. Another is that it should automatically > scale any existing magic layouts; currently, scalegrid will only > work right if you issue the command before you load or draw anything. > This past week, I created a standard cell library based on > something made by Tanner; their layout included a lot of half-lambda > geometry, and for that, the scalegrid feature worked like a charm; I > was able to read in the Tanner CIF and redo the layout in double the > units, with correct DRC. Probably the cifoutput generator needs to > be a little more sophisticated to deal with the contact squares. > We need to get the distribution out to a tar file on an ftp or > html server so people can start downloading it and testing it. Right > now it's just a CVS database on Rajit's system at Cornell. if his CVS server is accepting outside read-only connections, i should be able to "cvs update" it over. can you give me the site and directory name (and a good time to update, ie after you've checked in enough stuff). in addition to CVS, running a 'bugzilla'-like website (see http://bugzilla.mozilla.org/show_bug.cgi?id=33576 for an example) would be ultimately cool :-) > > Regards, > Tim
|
|