ftp.pasteur.org/FAQ//ftp-pasteur-org-FAQ.zip/FAQ/lsi-cad-faq/part1

Internet Message Format | 1997-01-10 | 43KB

Path: senator-bedfellow.mit.edu!bloom-beacon.mit.edu!howland.erols.net!worldnet.att.net!news.mathworks.com!newsfeed.internetmci.com!newsfeed.direct.ca!nntp.portal.ca!news.bc.net!info.ucla.edu!nnrp.info.ucla.edu!news.ucdavis.edu!altarrib!monk
From: [e-mail] (Michael Altarriba)
Newsgroups: comp.lsi,comp.lsi.cad,news.answers,comp.answers
Subject: comp.lsi.cad Frequently Asked Questions With Answers (Part 1/4) [LONG]
Supersedes: <lsi-cad-faq/[e-mail]>
Followup-To: comp.lsi.cad
Date: 10 Jan 1997 00:53:28 GMT
Organization: Department of Electrical and Computer Engineering, UC Davis
Lines: 970
Approved: [e-mail]
Distribution: world
Message-ID: <lsi-cad-faq/[e-mail]>
Reply-To: [e-mail]
NNTP-Posting-Host: monk.ece.ucdavis.edu
Summary: This is a biweekly posting of frequently asked questions with answers
         the for comp.lsi / comp.lsi.cad newsgroups. It should be consulted
         before posting questions to comp.lsi or comp.lsi.cad.
Keywords: FAQ
Xref: senator-bedfellow.mit.edu comp.lsi:7660 comp.lsi.cad:8482 news.answers:91629 comp.answers:23575

Archive-name: lsi-cad-faq/part1
Posting-Freqency: every 14 days
Url: http://www.ece.ucdavis.edu/sscrl/clcfaq/faq/faq-toc.html

Welcome to comp.lsi.cad / comp.lsi: this is the biweekly posting of fre-
quently asked questions with answers.  Before you post a question such as
"Where can I ftp spice from?", please make sure that the answer is not
already here.  If you spot an error, or if there is any information that
you think should be included, please send us a note at
[e-mail].

This FAQ has recently been put on the Web in a much more readable format.
Though it is still under minor construction, all of the pieces are there.
Try it out at <URL:http://www.ece.ucdavis.edu/sscrl/clcfaq/faq/faq-
toc.html> and let us know of any problems or suggestions by mailing to
[e-mail].

The products and packages described here are intended for research and edu-
cational use. As such, we try to limit our entries to applications which
are available for free or at low cost (< $500). We also wish to limit the
descriptions to at most a page (60 lines) in length.

    Bret Rothenberg <[e-mail]>
    Wes Hardaker <[e-mail]>
    Mike Altarriba <[e-mail]>

    Solid State Circuits Research Laboratory
    Electrical Engineering and Computer Science
    University of California, Davis
    Davis, California 95616

----------------------------------------------------------------------

  $Id: comp.lsi.cad.FAQ.ms,v 1.150 1997/01/10 00:52:03 altarrib Exp $

  Frequently Asked Questions with Answers

  ! 1: Mosis Users' Group (MUG)
    2: Improved spice listing from magic.
    3: Tips and tricks for magic (Version 6.3)
    4: What can I use to do good plots from magic/CIF?
    5: What tools are used to layout verification?
    6: EDIF data exchange format.
    7: What layout examples are available?
    8: How can I get my lsi design fabbed and how much will it cost?
    9: Mosis fabrication services.
    10: Archive sites for comp.lsi.cad and comp.lsi
    11: Other newsgroups and information sources that relate to comp.lsi*
    12: Simulation programs tips/tricks/bugs
    13: Getting the latest version of the FAQ
    14: Converting from/to GDSII/CIF/Magic
    15: CFI (CAD Framework Initiative Inc.)
    16: What synthesis systems are there?
    17: What free tools are there available, and what can they do?
    18: What Berkeley Tools are available for anonymous ftp?
    19: What Berkeley Tools are available through ILP?
  ! 20: Berkeley Spice (Current version 3f4)
    21: Octtools (Current version 5.1)
    22: Ptolemy (Current version 0.5)
    23: Lager (Current version 4.0)
    24: BLIS (Current version 2.0)
    25: COSMOS and BDD
    26: ITEM
    27: PADS logic/PADS PCB
    28: Another PCB Layout Package
    29: Magic (Current version 6.5)
    30: PSpice
    31: Esim
    32: iSPLICE3, a mixed-mode simulator for MOS/Bipolar circuits
    33: Watand
    34: Caltech VLSI CAD Tools
    35: Switcap2 (Current version 1.1)
    36: Test Software based on Abramovici text
    37: Atlanta and Soprano automatic test generators
    38: Olympus Synthesis System
    39: OASIS logic synthesis
    40: T-SpiceTM (was CAzM), a Spice-like table-based analog circuit simulator
    41: Galaxy CAD, integrated environment for digital design for Macintosh
    42: WireC graphical/procedural system for schematic information
    43: LateX circuit symbols for schematic generation
    44: Tanner Research Tools (Ledit and LVS) (Commercial Product)
    45: SIMIC, a full-featured logic verification simulator
    46: LASI CAD System, IC and device layout for IBM compatibles
    47: EEDRAW, an electrical/electronic diagramming tool for IBM compatibles
    48: MagiCAD, GaAs Gate Array Design through MOSIS
    49: XSPICE, extended version of Spice
    50: MISIM, a model-independent circuit simulation tool
    51: Nelsis Cad Framework
    52: APLAC, a general purpose circuit simulation and design tool
    53: SLS, a switch-level simulator
    54: OCEAN, a sea-of-gates design system
    55: ALLIANCE, a CAD package and simulator for teaching digital VLSI design
  ! 56: ceBox EDIF Viewer and Schematic Generator
    57: Analog CMOS VLSI Design Educational Resource Kit
    58: TDX Fault Simulation and Test Generation Software
    59: Nascent Technologies CDROM - magic and spice releases for Linux
    60: Time Crafter 1.0, a timing diagram documentation tool
    61: ACS, a general purpose mixed analog and digital circuit simulator
    62: LOG/iC, a logic synthesis package for PLDs
    63: SIMLAB, a circuit simulation environment
    64: Pcb, an X-based PC board design tool
    65: SPICE-PAC, A Modular Spice Simulator with Enhancements
    66: U.C. Berkeley Low-Power Cell Library
    67: The Substrate Resistance Extractor SUBSPACE
    68: XRLCAD, A C++ library for manipulating Calma (GDS) and CIF libraries
    69: SAVANT, an Analyzer of VHDL Applications for Next-Generation Technology
    70: Protel Demos for Windows
    71: BPECS PCB Software
    72: RF, an RF Circuit Simulation Tool
  + : new item
  ! : changed
  ? : additional information for this subject would be appreciated.

1: Mosis Users' Group (MUG)

  (From the Microelectronics Systems Newsletter)

  Microelectronic Systems News, formerly known as the MOSIS  Users' Group
  (MUG)  Newsletter,  includes not only items of interest to those design-
  ing integrated circuits for prototyping via MOSIS but also  for  those
  designing, prototyping and producing microelec- tronic systems. Notices
  of new items are broadcast to about  1800 subscribers  throughout  the
  world.  There is no charge for this service.

  To make a contribution or to be added to the  email  notification list,
  please send email to Prof. Don Bouldin at the University of Tennessee,
  Knoxville: [e-mail]

  Microelectronic Systems News can now be accessed at:

          <URL:http://microsys6.engr.utk.edu/ece/msn>

  A variety of design files and CAD tools contributed by the members of the
  MOSIS Users' Group (MUG) are now available via anonymous ftp from
  "ftp://ftp.mosis.edu/pub/mug" (128.9.0.32).  The files "readme" and
  "index" should be retrieved first.  These files are provided "as is", but
  may prove very helpful to those using the MOSIS integrated circuit proto-
  typing service.

2: Improved spice listing from magic.

  Hierarchical extractions with net names: ext2spice done by Andy Burstein
  <[e-mail]>:

  This program will do hierarchial extraction using node names.  It sup-
  ports PS, PD, AS, and AD extraction as well.  It is available for ftp
  from ftp://ic.eecs.berkeley.edu/pub/spice3/ext2spice.tar .

  Poly and well resistance extraction: There are persistent rumors that
  people have this working, however, all I have seen is extracted poly
  resistor with each end shorted together, ie each end has the same node
  name/number.

  (This is the most annoying problem that I typically encounter daily.  If
  ANYONE knows a fix for this, please tell us! I wrote a real quick and
  dirty set of scripts/programs to edit the magic file.  It will break the
  poly contacts and relabel them.  This is a real hack, but all other solu-
  tions require modification of the magic code itself.  This procedure only
  works with an extractor that handles labeled nodes, i.e. ext2spice from
  above.  --WH)

  Spice listing from magic with MESFETs.

  (from Jen-I Pi <[e-mail]>)

  We have a revised version (of sim2spice) that goes with version6. It is
  available from our anonymous FTP host
  "ftp://ftp.mosis.edu/pub/mosis/magic/gaas_extract.tar.Z" (128.9.0.32).

  Assuming file inv.ext exist, the procedure for using 'sim2spice' is

              ext2sim inv
              sim2spice inv.sim

  Here's the resulting SPICE decks for SPICE3e...

          SPICE 3 Deck created from inv.sim, tech=edgaas
          *
          z2 3 4 2 efet1.2 2.8
          C3 3 0    0.485F
          C4 4 0    1.062F
          z1 1 4 3 dfet1.2 2.8
          *

  Commercial Plotting Service

  Artwork Conversion offers an IC plotting service. We will take your
  CIF/GDSII files and plot them in large format color using an HP 650C
  color plotter in 24 hrs at a very reasonable cost.

  Designers can FTP files to artwork.com and we will plot them the same day
  and return by FedEx.

  Fill patterns and line types are completely customizable although most
  users select from the 100 preset patterns already defined.

  Complete information, specifications and pricing is available from our
  Web site: <URL:http://www.artwork.com/plot1.html>.

  We are offering universities our best price that we give to high volume
  customers: $4.00 per square foot of plotted area.

  The plots can be any size (the plotter is 36 inches wide).

  We can of course be reached by tel [phone] fax [phone].

  Foundries such as MOSIS, Chip Express and Orbit use this service as well
  as many small design shops that cannot justify a large format plotter.

  contact: Hagai Pettel [e-mail] or Steve DiBartolomeo
  [e-mail]

3: Tips and tricks for magic (Version 6.3)

  Searching for nets:

  Yes, magic does actually let you search for node names.  Use :specialopen
  netlist.  Then click on the box underneath label, you will be prompted
  for the name of the label you want to search for.  Enter the name, and
  then press enter twice.  Click on show, and then find, magic will then
  highlight the net.

  Bulk node extraction:

  Problems with getting the bulk node to extract correctly?  Try labeling
  the well with the node name that it is connected to.

  Painting Wells:

  Supposedly :cif in magic will automatically paint in the wells correctly.
  However this is not always the case.  If you are using mosis 2u technol-
  ogy, and your wells are getting strange notches in them, you might try
  changing the grow 300 shrink 300 lines in your lambda=1.0(pwell) and
  lambda=1.0(nwell) cif sections of your tech file to grow 450 shrink 450.
  (Remember you can use :cif see CWN to see nwell, if :cifostyle is nwell,
  or :cif see CWP to see pwell if its pwell technology to preview what will
  be done with the well.  You may use :feedback clear to erase what it
  shows you.)

  Magic notes available from ftp://gatekeeper.dec.com/pub/DEC/magic/notes
  (16.1.0.2):

  Magic note.1 - 9/14/90 - ANNOUNCEMENT:  Magic V6 is ready
  Magic note.2 - 9/19/90 - DOC:  Doc changes (fixed in releases after 9/20/90)
  Magic note.3 - 9/19/90 - GRAPHICS:  Mode problem (fixed 9/20/90)
  Magic note.4 - 9/19/90 - HPUX:  rindex macro for HPUX 7.0 and later
  Magic note.5 - 9/19/90 - GCC:  "gcc" with magic, one user's experience
  Magic note.6 - 9/19/90 - FTP:  Public FTP area for Magic notes
  Magic note.7 - 9/20/90 - RSIM:  Compiling rsim, one user's suggestions & hints
  Magic note.8 - 9/26/90 - GENERAL:  Magic tries to open bogus directories
  Magic note.9 - 9/26/90 - GRAPHICS:  Mods to X11Helper
  Magic note.10 - 10/5/90 - DOS:  Magic V4 for DOS and OS/2
  Magic note.11 - 10/11/90 - GENERAL:  reducing memory usage by 600k
  Magic note.12 - 12/19/90 - EXT2xxx:  fixes bogus resistances
  Magic note.13 - 12/19/90 - EXTRESIS:  fixed bug in resis that caused coredump.
  Magic note.14 - 12/19/90 - EXTRESIS:  new version of scmos.tech for extresis
  Magic note.15 - 12/19/90 - TECH:  documentation for contact line in tech file
  Magic note.16 - 12/19/90 - EXTRACT:  bug fix to transistor attributes
  Magic note.17 - 5/13/91 - CALMA:  Incorrect arrays in calma output
  Magic note.18 - 5/14/91 - CALMA:  Extension to calma input
  Magic note.19 - 6/28/91 - IRSIM:  Some .prm files for IRSIM
  Magic note.20 - 7/18/91 - EXTRESIS:  fixes for Magic's extresis command
  Magic note.21 - 2/7/92 - FAQ:  Frequently asked questions
  Magic note.22 - 11/6/91 - CALMA:  how to write a calma tape
  Magic note.23 - 11/4/91 - EXT2xxx:  fix for incorrect resistor extraction
  Magic note.24 - 11/8/91 - EXTRESIS:  fix 0-ohm resistors
  Magic note.25 - 11/15/91 - NEXT:  porting magic to the NeXT machine
  Magic note.26 - 11/21/91 - IRSIM:  fix for hanging :decay command
  Magic note.27 - 12/17/91 - RESIS:  fix for "Attempt to remove node ..." error
  Magic note.28 - 1/28/92 - MAGIC:  anonymous FTP now available
  Magic note.29 - 3/27/92 - PLOT:  support for Versatec 2700
  Magic note.30 - 4/8/92 - PATHS:  Have the ":source" command follow a path
  Magic note.31 - 4/10/92 - MPACK:  Mpack now works with Magic 6.3
  Magic note.32 - 3/13/92 - AED:  Using AED displays with Magic 6.3
  Magic note.33 - 3/13/92 - OPENWINDOWS:  Compilation for OpenWindows/X11
  Magic note.34 - 2/14/92 - OPENWINDOWS:  fix mouse problem
  Magic note.35 - 8/27/92 - RS6000: diffs to get magic to run on RS6000

4: What can I use to do good plots from magic/CIF?

  (Thanks to Douglas Yarrington <[e-mail]> and Harry
  Langenbacher <[e-mail]>, for feedback here.)

  CIF:

  CIF stands for CalTech Intermediate Form. It's a graphics language which
  can be used to describe integrated circuit layouts.

  (from Jeffrey C. Gealow <[e-mail]>)

  The definitive description of the Caltech Intermediate Form (CIF Version
  2.0) is included in Mead and Conway's book:

          @book{mead80,
            author = "Carver A. Mead and Lynn A. Conway",
            title = "Introduction to {VLSI} Systems",
            publisher = "Addison-Wesley",
            address = "Reading, Massachusetts",
            year = 1980,
            call = "TK7874.M37",

  A brief description is included in Rubin's book:

          @book{rubin87,
            author = "Steven M. Rubin",
            title = "Computer Aids for {VLSI} Design",
            publisher = "Addison-Wesley",
            address = "Reading, Massachusetts",
            year = 1987,
            call = "TK7874.R83",
            isbn = "0-201-05824-3"}

  Rubin's description should not be considered authoritative.  Parts of the
  description are not accurate.

  cif2ps  version 2 (Gordon W. Ross, MITRE):

  A much better version of cif2ps, extending the code of cif2ps (Marc
  Lesure, Arizona State University) and cifp (Arthur Simoneau, Aerospace
  Corp).  It features command line options for depth and formatting.  Can
  extend one plot over several pages (up to 5 by 5, or 25 pages). By
  default, uses a mixture of postscript gray fill and cross-hatching.
  Options include rotating the image, selecting the hierarchy depth to
  plot, and plotting style customization.  Plots are in B/W only.

  It was posted to comp.sources.misc, and is available by ftp from
  ftp://ftp.uu.net/usenet/comp.sources.misc/volume8/cif2ps.Z (192.48.96.9).

  cifplot:

  Cifplot plots CIF format files on a screen, printer or plotter.  Cifplot
  reads the .cif file, generates a b/w or color raster dump, and sends it
  to the printer.  Plots can be scaled, clipped, or rotated.  Hierarchy
  depth is selectable, as well as the choice of colormap or fill pattern.
  An option exists which will compress raster data to reduce the required
  disk space.  For those plotting to a Versatec plotter, there is also a
  printer filter/driver available called vdmp.

  oct2ps (available as part of the octtools distribution):

  It is possible to convert your .mag file to octtools, and then you may
  use oct2ps to print it.

  Both cif2ps and oct2ps work well for conversion to postscript.  They do
  look slightly different, so pick your favorite.  Note that cif2ps can be
  converted to adobe encapsulated postscript easily by adding a bounding
  box comment.  oct2ps does convert to color postscript, which can be a
  plus for those of you with color postscript printers.

  Flea:

  Flea ([F]un [L]oveable [E]ngineering [A]rtist) is a program used to plot
  magic and cif design files to various output devices. Parameters are
  passed to flea through the flags and flag data or through .flearc files
  and tech files.  Supports: HP7580 plotter, HP7550 hpgl file output,
  HP7550 plotter lpr output, Postscript file output, Laser Writer lpr out-
  put, Versatec versaplot random output.  Options include: Does line draw-
  ings with crosshatching for postscript, versatec, and hp plotters.  Many
  options (depth, label depth, scale, path, format...)

  Available by ftp from ftp://zeus.ee.msstate.edu/pub/flea.1.4.1.tar.Z .

  pplot:

  Can output color PostScript from CIF files. The source is available from:
  ftp://anise.ee.cornell.edu/pub/cad/pplot.tar.Z . It only generates PS
  files (including color PS), and there's no support for EPS files.  It is
  limited in its support of cif commands.  (Wire, roundflash, and delete
  are not supported.)  It only supports manhattan geometry (Polygons and
  rotations may only be in 90 degree multiples.)

  vic:

  Part of the U. of Washington's Northwest Lab, for Integrated Systems Cad
  Tool Release (previously UW/NW VLSI Consortium).  Does postscript and HP
  pen plotters.  Only available as part of the package.

  CIF/Magic -> EPS -> groff/latex

  Currently no prgram here directly generates EPS files.  It is possible to
  add an EPS bounding box (%% BoundingBox: l t b r) to the output from
  these programs to get an EPS file.  Alternatively, ps2eps or ps2epsf may
  be used.

  CIF display on PCs

  LaSy

  (from Frank Bauernoeppel <[e-mail]>)

  The primary goal of LaSy was to implement a simple CIF layout viewer
  under MS-Windows.

Requirements:
  MS-Windows 3.1 in extended mode or Windows-NT. Hi-resolution colour
  display, mouse, and a colour-printer are recommended. Note that there is
  a special Windows-NT version of LaSy included: lasy32.exe featuring 32-
  bit integer coordinates.

Input:
  A CIF file plus appropriate layer description (.lay file). Sample layer
  descriptions are included.  You probably have to adapt them to your tech-
  nology.  CIF description see "Introduction to VLSI systems" by Mead and
  Conway.  Several restrictions apply (cf. online help).

Output:
  A layout window for visual inspection/measurements of the layout.
  Printer output using Windows printing mechanism, works fine.  Clipboard
  copy in bitmap and metafile format. The metafile is a flat, object
  oriented layout representation understood by many applications. Can be
  postprocessed with MSDraw among others.  The bitmap gives a pixel
  oriented view of the layout (at screen resolution) and can be postpro-
  cessed by most "Painting programs".

  I have repacked the archive for distribution (some designs removed).

  The new url is: <URL:ftp://ftp.informatik.hu-
  berlin.de/pub/local/hulda/lasy25.zip>

  Two references that describe the CIF file formats are:

          Introduction to VLSI Systems, Mead & Conway, 1980, p115
          and
          Basic VLSI Design, Pucknell & Eshraghian, 1988, p 275

5: What tools are used to layout verification?

  Gemini:

  Gemini is a graph isomorphism tool for comparing circuit wirelists.  The
  latest version of Gemini is 2.7 and is now available by FTP from
  shrimp.cs.washington.edu (128.95.1.99).  Note: Gemini is not available by
  anonymous FTP.  Send email to Larry McMurchie ([e-mail])
  if you need the FTP login and password for Gemini.

  Version 2.7 includes a new SIM file format to support four-terminal MOS
  transistors.  This format is called 'LBL' and was inspired by Mario
  Aranha at Lawrence Berkeley Labs.  Also some minor bugs have been fixed
  concerning portability.  The user guide 'gemuser.ps' has been updated to
  reflect the changes to the code.

  Gemini compiles and runs on a wide variety of architectures, including
  Sparc, Mips, DEC AXP, HP, KSR, Intel i860, MC 68020 and VAX, under both
  Classic C and ANSI C compliant compilers.  As the number of architectures
  continues to expand, new portability problems are revealed.  Please keep
  us informed if you encounter any portability problems or bugs.

  Contact:

          Larry McMurchie
          Computer Science Department, FR-35
          University of Washington
          Seattle, WA  98195
          [e-mail]

  Tanner LVS:

  This is a relatively inexpensive commercial product, see the section on
  Tanner tools.

  Wellchecker:

  (from MUG) ftp ftp.mosis.edu (128.9.0.32)

  netcmp:

  Part of the caltech tools (see the "Caltech VLSI CAD Tools" section)

6: EDIF data exchange format.

  (from Mark Lambert <[e-mail]>)

  The Electronic Design Interchange Format (EDIF) is the most widely used
  EDA standard and is used to interchange design data between CAD systems.

  The language is a standard under the auspices of the Electronic Indus-
  tries Association (the `EIA'), a US based industry association, responsi-
  ble for a number of electronics related standards. EDIF Version 3 0 0,
  used for the transfer of connectivity and schematic information, has also
  become an IEC standard; IEC 1690.

  The latest version of EDIF, Version 4 0 0, promises to add to EDIF Ver-
  sion 3 0 0 in the areas of; PCB and MCM Capabilities, Technology Rules
  and Manufacturing Drawings

  EDIF Version 4 0 0 is currently out for ballot, until 28th May 1996, as
  EDIF Version 3 9 9.

  EDIF Version 3 9 9 documentation is supplied in CD-ROM form or on paper
  directly from the EIA or Framemaker4 hypertext format directly from the
  EDIF Technical Centre, on behalf of the EIA. To review the material, the
  FrameViewer software is required. To obtain a copy of EDIF Version 3 9 9
  (4 0 0), contact either Patti Rusher of the EIA or the EDIF Technical
  Centre for detailed instructions.

  The EDIF Version 2 0 0 Reference Manual and User Guides and copies of the
  manuals or CD-ROM for EDIF Version 3 0 0 can be obtained from the Elec-
  tronic Industries Association, Attn. Patti Rusher.

  For more EDIF related information visit the EDIF Web site:

          http://www.edif.org/

  and the anonymous ftp server:

          ftp://edif.cs.man.ac.uk/pub/edif

  An ftpmail server is provided for those without ftp access. Send an empty
  email message to: [e-mail] ; a message describing the com-
  mands which can be used in further email messages to retreive files will
  be sent to you.

  An electonic mailing list is available to people interested in EDIF and
  for EDIF developers/programmers. Send email to edif-users-
  [e-mail] to be added.

          Patti Rusher at the EIA can be contacted at:
               Patti Rusher
               2500 Wilson Boulevard, Suite 203
               Arlington, VA 22201, U.S.A.

               Telephone: [phone]
               Fax: [phone]
               E-mail: [e-mail]

          The EDIF Technical Centre can be contacted at:
               EDIF Technical Centre, Department of Computer Science,
               University of Manchester, Manchester M13 9PL, UK

               Tel: [phone]
               FAX: [phone]
               E-mail: [e-mail]
               URL: http://www.cs.man.ac.uk/cad/EDIFTechnicalCentre/

7: What layout examples are available?

  From MUG:

  Analog neural network library of cells, 66-bit Manchester carry-skip
  adder, static ram fabricated at 2-micron, an analog op amp, from
  ftp://ftp.mosis.edu/pub/mug .

8: How can I get my lsi design fabbed and how much will it cost?

  See section on mosis fabrication services as well.

  (From MUG 20 George Lewicki of Orbit Semiconductor)

  Orbit Semiconductor operates an integrated circuit prototyping service
  that accepts designs each week for all of its processes.  The service is
  available to both U.S. and non-U.S. designers. In- quiries about the
  FORESIGHT prototyping service should be ad- dressed to George Lewicki.
  Designs can now be submitted directly via email.

              Orbit Semiconductor, Inc.
              1215 Bordeaux Drive
              Sunnyvale, CA 94089
              TEL: [phone]
              FAX: [phone]
              Email: [e-mail]

  (Contributed by Don Bouldin of the University of Tennessee)

  Recently, I contacted several foundries to determine  which  com- panies
  are  interested  in fabricating small to moderate lots of wafers for cus-
  tom CMOS designs.  I believe many of the readers of this  column are
  designers who wish to have fabricated only 1,000 to 20,000 parts per
  year.  There are currently several  prototyp- ing  services  (e.g. MOSIS
  and Orbit) that can produce fewer than 100 parts for about $100 each and
  there are  also  several  foun- dries  which  are willing to produce
  100,000 custom parts for $5- $20 each (depending on the die size and
  yield).  My  purpose  was to  identify  those companies filling the large
  gap between these two services.

  The prices in the table below are a result of averaging the  data sup-
  plied by four foundries.  The raw data varied by more than +/- 40% so the
  information should be used only in the early stages of budgetary  plan-
  ning.   Once  the design specifications are fairly well known, the
  designer should contact one or more foundries  to obtain  specific
  budgetary  quotes.  As the design nears comple- tion, binding quotes can
  then be obtained.

  The following assumptions were made by the foundries:

  All designs will require custom CMOS wafer  fabrication  using  a
  double-metal, single-poly process with a feature size between 2.0 and 1.2
  microns.  The designs may contain some  analog  circuitry and  some  RAM
  so the yield has been calculated pessimistically.  The dies will be pack-
  aged and tested at 1  MHz  using  a  Sentry- type digital tester for 5-10
  seconds per part.  The customer will furnish the test vectors.

          Piece Price includes Wafer Fabrication+Die Packaging+Part Testing
          Size        Package                      Quantity

                                 |1,000 | 5,000 | 10,000 | 20,000  |100,000
          -----------------------------------------------------------------
          2 mm x 2 mm; 84 PLCC:  | $ 27 | $  6  |  $  5  |  $  4   | $  3 |
          5 mm x 5 mm; 84 PLCC:  | $ 31 | $ 12  |  $  8  |  $  7   | $  6 |
          5 mm x 5 mm; 132 PGA:  | $ 49 | $ 30  |  $ 25  |  $ 22   | $ 18 |
          7 mm x 7 mm; 132 PGA:  | $ 65 | $ 44  |  $ 36  |  $ 31   | $ 27 |

          Lithography charges:  $ 20,000 - $ 40,000
          Preferred Formats:  GDS-II or  CIF Tapes
          Additional charges for Second-Poly:  $ 5,000

  (This is from MUG 19, there is also a list of foundries that these prices
  were derived from.  In the interested of saving space, I have ommitted
  the list.  The list is available from MUG's ftp site included in MUG
  newsletter #19.)

9: Mosis fabrication services.

  (From Mosis) Information is available from mosis for pricing and fab
  schedules through an automatic email system:

  Mail to [e-mail] with the message body as follows:

          REQUEST: INFORMATION
          TOPIC: TOPICS
          REQUEST: END

  for general information and a list of available topics.

  If you need to contact a person at mosis, you may mail to [e-mail]
  with REQUEST: ATTENTION.

  Also anonymous ftp is available. ftp to ftp.mosis.edu.  This is a dupli-
  cation of all files that are available from the mail server.

  (From MUG 20 Contributed by Don Bouldin of the University of Tennessee)

  Multi-project fabrication of BICMOS designs are already available to
  European universities via CMP and to Canadian universities via the Cana-
  dian Microelectronic Corporation.  However, in the United States, the
  demand for BiCMOS fabrication via MOSIS has not been considered signifi-
  cant.  MOSIS is currently planning to start offering 0.5-micron BiCMOS
  during the first quarter of 1994. This will have a core voltage operation
  of 3.3v and a clock frequency in the range of 220-250Mhz.  MOSIS is
  interested in seeing if a larger demand exists in the community than
  expressed so far.

  If you would like to have BiCMOS available before 1994, please send a
  short note to [e-mail] (with a copy to [e-mail])
  using the following format.

               REQUEST:  ATTENTION
                          .
                          .
                   your message goes here
                          .
                          .
               REQUEST: END

  (From MUG 20 and Chris Donham of the University of Pennsylvania)

  Support for mosis technologies under Cadence Analog Artist 2.4 is avail-
  able as is from University of Pennsylvania.  This includes DRC, LVS, EXT,
  and a beginner's guide.  Currently they are working on support for Opus
  4.2.  The files supporting Artist 2.4 are currently available via
  anonymous FTP.  Penn is not affiliated with MOSIS, except as a satisfied
  customer, and as a result, NO WARRANTY IS EXPRESSED OR IMPLIED WITH
  REGARDS TO THE FILES, OR THEIR FITNESS FOR ANY USE.  Use the files at
  your own risk.  To obtain the files, FTP to axon.ee.upenn.edu
  (130.91.6.208), using the name "anonymous" and your mailing address as
  the password.  The files are in the "pub" directory.

  Penn is in the process of switching from Artist 2.4 to Opus 4.2.  The
  manual is being rewritten, and the support files are being updated.
  Technology files supporting DRC, Extract, and Compare are currently in
  beta-test.  If problems or bugs are detected, please send email to
  "[e-mail]".

10: Archive sites for comp.lsi.cad and comp.lsi

  (None of these are comprehensive archives, rather, they have about 3
  postings each)

  comp.lsi.cad:
  ftp://cnam.cnam.fr/pub/Archives/comp.archives/auto/comp.lsi.cad
  ftp://cs.dal.ca/pub/comp.archives/comp.lsi.cad
  ftp://srawgw.sra.co.jp/.a/sranha-bp/arch/arch/comp.archives/auto/comp.lsi.cad

11: Other newsgroups and information sources that relate to comp.lsi*

  alt.cad
  comp.cad.cadence
  comp.lang.verilog
  comp.lang.vhdl
  comp.sys.mentor
  sci.electronics

  The following gopher link points to a collection of information from
  pulled from newsgroups like comp.lsi.cad, comp.lsi, and other cad related
  sources.

  gopher://kona.ee.pit.edu/

12: Simulation programs tips/tricks/bugs

  Berkeley spice:

  Pspice:

  Hspice:

  If your simulation won't converge for a given DC input, you can ramp the
  input and print the DC operating point and then set the nodes that way
  for future simulations.

  A number of documents are available for information on BSIM model parame-
  ters: (from Mark Johnson, as posted to comp.lsi <[e-mail]>)

  1. The very best written description I have seen is in a software manual.
     The good news is that this manual is free; the bad news is that you
     have to buy the multi-thousand-dollar program in order to get the free
     manual.  The program is HSPICE from Meta-Software Inc (Campbell,
     Calif., USA).  The HSPICE User's Manual, chapter 7, gives all the
     details you'd ever want to know regarding BSIM parameters.

  2. The second best description I have seen of BSIM is in, strangely
     enough, a manual for BSIM2 (!).  It is available from the University
     of California at Berkeley.  Telephone [phone] and they will
     give you instructions on how to buy the manual.  (They'll probably
     suggest that you might want to buy some software too).

             J.S. Duster, M.C. Jeng, P.K. Ko, and C. Hu, "Users
             Guide for the BSIM2 Parameter Extraction Program and
             the SPICE3 with BSIM Implementation"

  3. You can learn some things about BSIM parameters by reading about pro-
     grams which extract the parameters from measured data.  UC Berkeley
     offers several programs and manuals for this.  The one that I person-
     ally prefer is

             M.C. Jeng, B.J. Sheu, and P.K. Ko: "BSIM Parameter
             Extraction - Algorithms and User's Guide," Memo
             No. UCB/ERL M85/79, 7 October 1985.

  4. Next, look at Sheu's Ph.D. thesis.  He is the guy who combined the
     Bell Labs CSIM model with a bunch of other published equations, and
     formulated BSIM.  It's available from the same phone number.

             B.J. Sheu, "MOS Transistor Modelling and Characterization
             for Circuit Simulation", Memo No. UCB/ERL M85/85,
             26 October 1985

  5. The worst description (in +my+ opinion of course) is unfortunately in
     the most-accessible publication.  To save space in the journal they
     left out some parameter discussions and (again in my opinion) produced
     a disjointed, not-fully- informative paper.  Others may have different
     views, naturally.

             B.J. Sheu, D.L. Scharfetter, P-K Ko, M-C Jeng, "BSIM:
             Berkeley Short-Channel IGFET Model for MOS Transistors,"
             IEEE Journal of Solid-State Circuits, Vol SC-22, No. 4,
             August 1987, pp. 558-565.

13: Getting the latest version of the FAQ:

  Mail to [e-mail] with the subject "send faq".

  If you wish to be added to the FAQ mailing list, send a note to lsi-faq-
  [e-mail] with subject heading 'subscribe'. You will then
  have the FAQ regularly emailed to the return address of the note. Like-
  wise, use the subject heading 'unsubscribe' to be removed from the list.

  This FAQ is now cross-posted to news.answers and comp.answers. This news-
  group is archived periodically on
  ftp://rtfm.mit.edu/pub/usenet/news.answers/lsi-cad-faq [18.181.0.24].
  Postings are archived as "part1" through "part4".

  Our FAQ is also available through the WWW pages.  You can access it at
  <URL:http://www.ece.ucdavis.edu/sscrl/clcfaq/faq/faq-toc.html> .  I sug-
  gest this site above the one listed below, since ours is hyper-text for-
  matted and the site below is essentially just a text to html conversion
  with no table of contents.

  (from Thomas A. Fine <[e-mail]>)

  WWW I maintain an "archive" of news.answers available via WWW.  As a
  matter of fact, I used WWW to read through your posting just last week.
  I found it very informative; thanks much.  Advertise the following refer-
  ence to get to the archive in general:
    <URL:http://www.cis.ohio-state.edu:80/hypertext/faq/usenet/FAQ-
  List.html>

  or to get to your particular FAQ, give out this reference:
    <URL:http://www.cis.ohio-state.edu:80/hypertext/faq/usenet/lsi-cad-

  faq/top.html>

  Gopher The news.answers introduction (which I pulled up in WWW ;-) lists
  the following gopher sites for the FAQs:

              cc1.kuleuven.ac.be port 70
              jupiter.sun.csd.unb.ca port 70
              gopher.univ-lyon1.fr, port 70
              ftp.win.tue.nl, port 70
              gopher.win.tue.nl, port 70
              kona.ee.pitt.edu 70

  To reference gopher from Mosaic, us the following reference:
    gopher://kona.ee.pitt.edu WAIS

  I pulled this straight out of the news.answers Introduction:

  Note that the periodic posting archives on rtfm.mit.edu are also accessi-
  ble via WAIS (the database name is "usenet" on port 210).  If you don't
  know what WAIS is, don't worry about it, although you can look in
  comp.infosystems.wais if you're curious.  And don't write to us and ask,
  please; we unfortuately already have too many things to deal with without
  having to answer questions about other people's software.

14: Converting from/to GDSII/CIF/Magic

  Magic version 6.3 is capable of reading and writting to all three for-
  mats.  (From the magic man page):

  calma [option] [args]

  This command is used to read and write files in Calma GDS II Stream for-
  mat (version 3.0, corresponding to GDS II Release 5.1).  This format is
  like CIF, in that it describes physical mask layers instead of Magic
  layers.  In fact, the technology file specifies a correspondence between
  CIF and Calma layers.  The current CIF output style (see cif ostyle) con-
  trols how Calma stream layers are generated from Magic layers.

  (from Jeffrey C. Gealow <[e-mail]>)

  Calma Company sold their electronics CAD/CAM software (GDS II) to Valid
  Logic Systems which later merged with Cadence.

  Cadence has added a few extensions.  A Cadence document is almost identi-
  cal to the old Calma Company document:

          Cadence Design Systems, Inc.

          Construct Stream Format
          Reference

          Version 4.0
          August 1991

          900-001094

  An overview of the Stream format is included in Rubin's book:

          @book{rubin87,
            author = "Steven M. Rubin",
            title = "Computer Aids for {VLSI} Design",
            publisher = "Addison-Wesley",
            address = "Reading, Massachusetts",
            year = 1987,
            call = "TK7874.R83",
            isbn = "0-201-05824-3"}

  cif [option] [args]

  Read or write files in Caltech Intermediate Form (CIF).

15: CFI (CAD Framework Initiative Inc.)

  (From Randy Kirchhof <[e-mail]>)

              CFI quick FAQ guide for release 1.0, v1.1

  For those of you who may be unfamiliar with our work, The CAD Framework
  Initiative Inc. was formed in May 1988. We're located in Austin, TX,
  although we're a distributed company. We're a  not-for-profit consortium
  formed under the laws of the state of Delaware.  Our mission is to pro-
  vide industry-accepted standards and technology that enable interopera-
  bility of electronic design automation (EDA) applications and data for
  end-users and suppliers world-wide.  This includes interoperability
  between EDA applications as well as the integration of EDA applications
  into CAD frameworks.

  A CAD framework is a software infrastructure which provides a common
  operating environment for CAD tools.  Through a framework, a user should
  be able to launch and manage tools, create, organize, and manage data,
  graphically view the entire design process and perform design management
  tasks such as configuration management, version management, etc.  CFI
  Release 1.0 started shipping in January 1993.

  Q      When can users buy CFI compliant tools?

  A      Eleven vendor companies have announced EDA products and frameworks
         which will be available and compliant with CFI 1.0 standards. CFI
         has initiated a formal certification program for these (and future
         products) as of 12/93. CFI expects to begin awarding the first
         certification brand marks in the first quarter of 1994.  We expect
         to see a rapid expansion of compliant products beginning in the
         third quarter of 1994.

  Q      How can the Standards be obtained?  Are there any restrictions?

  A      The 1.0 Standards, copyrighted by CFI, are available to members
         and non-members priced as a set or individually through CFI Member
         Services [phone].  They will also being distributed under
         license by Cadence, Mentor Graphics, and Viewlogic as part of
         their product documentation.  Versions of the 1.0 Standards are
         available on diskette in an electronic format as well as bound
         manuals.

  Q      How do the CFI Standards relate to vendor framework programs like
         Mentor's Open Door, Viewlogic Power Team and Cadence Connection
         Partners - with so many point tool vendors participating, don't
         they have this problem solved?

  A      The major EDA vendors have been and continue to be challenged by
         their customers over multi-vendor integration.  These programs
         were a practical response by opening up their existing interfaces
         and providing services to assist integration.  CFI 1.0, and future
         releases, will create a functional alternative to a growing subset
         of those interfaces so that the requirement that point tool ven-
         dors create partnership specific versions of their tool will
         decrease.  Actually, the service provided through these programs
         will likely compliment the CFI certification effort as these
         supplier's frameworks become fully certified.

Contact: [e-mail] (CFI Member Services, Jean Gallagher) CFI Main number:
[phone]   Fax: [phone]

16: What synthesis systems are there?

  Thanks to Simon Leung <[e-mail]>, Michel Berkelaar
  <[e-mail]>, Noritake Yonezawa <[e-mail]>, Donald A
  Lobo <[e-mail]>, Greg Ward <[e-mail]>, Peter Duzy,
  Robert Walker <[e-mail]>, Heinrich Kraemer
  <[e-mail]>, Luciano Lavagno <[e-mail]>

  ADPS
  - Case Western Reserve University, USA
  - scheduling and data path allocation
  - Papachristou, C.A. et al.: "A Linear Program Driven Scheduling and
    Allocation Method Followed by an Interconnect Optimization Algorithm",
    Proc. of the 27th DAC, pp. 77-83, June 1990.

  ALPS/LYRA/ARYL
  - Tsing Hua University
  - scheduling and data path allocation
  - Lee, J-H: et al.: "A New Integer Linear Programming Formulation of
    the Scheduling Problem in Data Path Synthesis", Proc. of ICCAD89, pp.
    20-23, November 1989.

  BDSYN
  - University of California, Berkeley, USA
  - FSM synthesis from DECSIM language for multilevel combination-logic
    realization
  - Brayton, R.: "Multiple-level Logic Optimization System",  Proc. of IEEE
    ICCAD, Santa Clara, Nov. 1986

  BECOME
  - AT & T Bell Labs, USA
  - FSM synthesis from C-like language for PLA, PLD and standard cell realization
  - Wei, R-S.: "BECOME: Behavior Level Circuit Synthesis Based on Structure
    Mapping", Proc. of 25th ACM/IEEE Design Automation Conference, pp. 409-414,
    IEEE, 1988

  BOLD
  - logic optimization
  - Bartlett, K. "Synthesis and Optimization of Multilevel Logic Under Timing
    Constraints", IEEE Transactions on Computer-Aided Design, Vol 5, No 10,
    October 1986

  BRIDGE
  - AT & T Bell Labs, USA
  - High-level synthesis FDL2-language descriptions
  - Tseng: "Bridge: A Versatile Behavioral Synthesis System", Proc. of 25th
    ACM/IEEE Design Automation Conference, pp. 415-420, IEEE, 1988