GDSreader - simple Calma (GDSii) parser/printer tool.

This software has as target the printing/plotting/displaying of Calma (GDSii)
files without using true layout editors. I had once to visualize an unknown
Calma file and customizing LEdit or Magic (the two layout editors I had access
to) was so difficult that I decided to write this program.

License
gdsreader is placed under the GNU Public License (GPL) version 2.0. See the
COPYING file for more information.

Current status:
- gdsreader is in an alpha stage and you should not expect too much from it;
- the Calma files are almost completely parsed (had no layout example that makes
  use of BOX/NODE elements);
- given a Calma structure name, a PostScript file and a HPGL/2 file are
  generated. The way each layer is handled is controlled by an ASCII
  configuration file. The properties that can be set are color, fill (only
  solid is supported), hatch (simple or cross, the angle and spacing are user
  customizable too).

Installation
Take a look at the Makefile and change it accordingly to your setup. Type make.
In order to produce an useful PostScript output, you need to write a
configuration file (default is .layers.config). The one you'll find with the
distribution is suitable for the Calma example test.gds (an actual Bandgap
reference).

What needs to be done:
- to try a different way to do hatching in PostScript (e.g. by using custom
  patterns);
- to do shape merging for a better output;
- to rewrite parts of the current C code in order to make it more efficient
  (there are lots of repeating loops for seeking same things, linear search
  might be replaced by more clever ways). Up to now, the code was written only
  to work, the efficiency was not a major concern;
- an unified rc file, for setting all aspects of the printing/plotting (page
  size etc.);
- documentation etc.
Any other ideas are welcome.

For suggestions, comments or bug reports, please drop an email to:
serbanp@ix.netcom.com

Thanks
Have to thank to Steve Tell (tell@cs.unc.edu) for its stream_utils-0.10 package,
useful as an example of how to deal with the Calma binary file, and to various
people who have been so kind to reply my posts in the comp.cad.lsi and
comp.lang.postscript newsgroups.


Good luck!

New feature for gdsreader-0.3.2

  Generation of POVRAY scene files.

Quick Overview:

GDS2, (or "Calma"), files contain 2D shape data, and each shape has
a specific layer number, (gdsno), and a datatype number.  To  make
any sort of 3D model, these two integer numbers must be  "mapped"
to a corresponding thickness and depth parameter, and  also to a
color specification.   This  is handled by the user-created
"layers.config" file.  A  layers.config file needs to be made for
each technology you wish to convert.  A process which differs only
in the number of metal layers used is considered a different
technology.  The layers.config file is also used in the generation
of postscript and HPGL output, so several of the parameters in
a layers.config file do not necessarily apply to POVRAY output.

The layers.config file you make must have entries of the form:

newlayer
 gdsno 19
 hidden no
 name m1
 red 0.4
 green 0.2
 blue 0.8
 ...
 thickness 0.630000
 depth 2.530000
endlayer

...for each layer you want to have visible in the POVRAY file.

You can see the list of layers used in the IBM design kits by
examining the cds2gds.map file.  It contains lines with layer
name, purpose, layer number, and datatype number.

A utility called "maptolayer" is included which can be used to
generate a dummy layers.config template from a cds2gds.map file
from a cadence kit.

Example:
  ./maptolayer < cds2gds.map > mylayers.config

The dummy layers file has unique colors for each layer as well
as depth and thickness.  You must edit these to unhide  ("hidden
no") the desired layers, and set correct depth and thickness for
them.  The depth is the height of the top of the layer in
microns, and  the thickness specifies how much further "down" it
goes.  This utility IGNORES the datatype information in the map
file, and this may require tweaks on the part of the user to fix.

In many GDS2 files, each metal layer uses a unique gdsno number to
identify it, and the datatype number can be ignored.  On others
however, two different metal layers may use the same gdsno, but a
different datatype number.  To handle this, by default a shape is
mapped to a specific gdsno and datatype if there is an entry for
it in the layers.config file.  If there is no match, an entry that
just specifies the gdsno is used.  You may need to add complete
newlayer...endlayer sections for layers that share gdsno values.

This can also be used if you  want to make objects look different
even if they are on the  same physical level.  For example, by
using the datatype  parameter, you might have vias be a different
color than viabars.

Sample files "layers.config.5AM.4ML" and "layers.config.7HP.5ML"
are included.  The 5AM file does not use datatypes, but the
7HP one does.  IBM may change the mapping with each kit revision,
but hopefully they will not.

The colors used in the povray output are those supplied in the
layers.config file.  Hatch/fill is ignored for povray output.  You
can generate  a povray file of the cell "topcell" in a gds2 file
"mydesign.gds2" using the command:

 gdsreader -i mydesign.gds2 -c mylayers.config -s topcell

By default output will be stored in the file "result.pov". A
postscript file will be called result.ps, etc.  You can select
any subcells in the gds2 file by giving it's name instead of
the "topcell".

Other  command line options can be used to change the default
output file names. The POV scene file is complete with a camera
and light source, but they  should be repositioned/adjusted to get
a good  view.  A sample command to  quickly render the scene might
be:

 povray +P +W640 +H480 +Q2 -I result.pov
   (note: This only works on new versions of povray, see
     below for an alternate command.)

This will display an image on the screen.  Click on it to close
the window.  Read the documentation on POVRAY before trying to
edit  the .pov file.  The following will render the same image to
a file called 'result.ppm' in ppm format:

 povray -D0 +W640 +H480 +Q2 -I result.pov +FP

You can view this with 'xv result.ppm'

A much nicer picture or movie can be created using shadows and
anti-aliasing options but that is beyond the scope of this
README file.  Consult the POVRAY documentation first.

Note that the top cell object is instanced at the very bottom
of the pov file and it is usually scaled so that the micro-based
vertical dimensions are proportional to the integer-grid based
layout dimensions.

You might wish to add a plane representing the substrate by
appending something like:

plane {
 <0,1,0>, -30
 texture { pigment { color rgb <0.400000,0.400000,0.400000> } }
}


...to the end of the POVRAY file.

  Good luck with this...

     - Pete