[Gnucap-devel] Re: gEDA: GTKWave release 2.0.0pre3

[Al Davis]

On Friday 07 February 2003 09:44 am, Burnie West wrote:
> How much sim data?  This is a highly user-dependent kind of question.
> For my own purposes, I normally use HSPICE with a .OPTIONS PLOT=2
> "card" -- this dumps all active nodes and creates a HUGE data file.

> Then, I will postprocess this data file to create a tab-delimited
> data file containing whichever nodes I'm interested in, one that
> can be read by any spreadsheet program.  A question for this group:
> would the tools I've built for loading, comparing, combining (*, +, /)
> edge-detecting, peak-detecting, and time-displacement comparison of
> transient waveforms be useful to you? 

Yes!!!!!
GPL?

> Part of the answer needs to
> comprehend that the input to these tools is one or more HSPICE .trn
> files (created by the above option and not available in gnucap).

This problem is easy to solve.

I have been planning to add a way to do a large dump for post-processors, for 
years.  It has never made it to the top of the to-do list.  The existence of 
a tool to make use of it can change that quickly.

Gnucap can probe a lot more internal points than Spice can.  Actually saving 
everything could produce data for hundreds of probes in a trivial circuit.  A 
select mechansim similar to the plot or print command would take care of it.

The next question is the file format ...

Gnucap uses a nice white-space delimited format now, probably compatible with 
that tab delimited file you make now.

The spice raw format is common and probably should be supported, but it is 
bigger and harder to read than the format gnucap already uses.  Its only 
advantage is spice compatibility.

There is a need for a more efficient format for this, but it can come later.

I cannot find any documentation on the .trn format.