[Magic-dev] TSMC .18

[Jeff Sondeen]

sorry for the confusion, but the reason that techfile you looked at is
virtually empty is "hidden" in the CHANGELOG file:

  ftp://ftp.isi.edu/pub/sondeen/magic/new/beta/CHANGELOG

 also, some
       of the full verions *.TSMC.tech27 techfiles are completely empty
              -- only the "light" versions are available (see discussion
       below under "2001-05-?? release 2001a")
...

      note that the SCN6M_DEEP.09.TSMC.tech27 and
      SCN6M_SUBM.10.TSMC.tech27 regular versions could not be built (as
      they have too many types) but that their 'light' versions (see
      below) are present: SCN6M_DEEP.09.TSMC.light.tech27 and
      SCN6M_SUBM.10.TSMC.light.tech27


((actually the above sentence is wrong, only the non-.TSMC versions
(ie., withOUT HV transistor layers) are there:

 pub/sondeen/magic/new/beta/2002a/SCN6M_DEEP.09.light.tech27
 pub/sondeen/magic/new/beta/2002a/SCN6M_SUBM.10.light.tech27

((you probably want the 2002a version)):


 ...


   *.light.tech27

     added "light" versions of all techfiles -- these techfiles omit
     many of the rarely used layers of the techfiles, namely "fill",
     "pins", "resistor", and implant layers (many of these special
     layers are described in later in this file).


Anyway, back to the problem is missing HV layers, all of the SCN6M
techfiles have too many types to allow all the extra types that HV
layers create, altho a lot of the types are taken by contacts (such as
'pc') in the 'contact' section of the techfile, which historically
took 3 (hidden) types per actual symbolic contact type (for the
"images"), but i believe that's one of the things that Tim fixed in
magic version 7.3.

once you get 7.3 running, i'd be glad to give you a real
SCN6M_DEEP.09.TSMC.tech27 (or more likely
SCN6M_DEEP.09.TSMC.light.tech27) file to see if it can be read into
the new magic.  you still may have to manually delete some types you
don't plan on using (like 'mim', etc). ((i still don't have 7.3
running on solaris -- no time to track down the compile problems. i
should be getting a linux machine before in the next couple of months
so can get back to techfiles then)).

/jeff


Chris Fearing writes:
 > I am considering using TSMC .18 technology for my next chip.  I have 
 > previously used TSMC .25 with magic 7.1
 > 
 > In opening up magic with the technology file SCN6M_DEEP.09.TSMC.tech27 
 > from the 2001a library, I get these errors
 > 
 > Section "styles" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "contact" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "compose" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "connect" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "cifoutput" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "cifinput" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "mzrouter" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "drc" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "extract" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "wiring" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "router" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "plowing" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Section "plot" was missing from 
 > ./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27.
 > Cannot load technology "./cad/lib/2001a/SCN6M_DEEP.09.TSMC.tech27"
 > 
 > and magic does not open.  I do not get these errors using tech file 
 > SCN6M_DEEP.09.tech27
 > 
 > There is not a SCN6M_DEEP.09.TSMC.tech27 in the 2002a library.
 > It also appears that the hv layers used in pads are not defined in this 
 > tech file.
 > 
 > 
 > Do I need to upgrade magic versions(7.2 or 7.3) to work with a .18 
 > technology?
 > Can I use SCN6M_DEEP.09.tech27 for TSMC .09.
 > Has anyone created pads for this technology?
 > Are there any other caveats to the combination of magic and TSMC .18 
 > that I am missing?
 > 
 > Thanks
 > 
 > Chris Fearing
 > 
 > _______________________________________________
 > magic-dev mailing list
 > magic-dev at vlsi.cornell.edu
 > http://vlsi.csl.cornell.edu/mailman/listinfo/magic-dev