Re: gEDA-user: Simulation troubles ...

[Dan McMahill <dan@xxxxxxxxxxxx>]

Kai-Martin Knaak wrote:
> On Tue, 06 Nov 2007 20:30:14 -0500, al davis wrote:
> 
>> A serious viewer for analog simulation data can plot any data against
>> any other, show in the S plane, show triggered data such as "eye"
>> diagrams, do math on the waveforms, overlay waveforms scaled and time
>> shifted, ......  And ... it has an extension language.  It also
>> supports, or at least doesn't interfere with, the scripting ability of
>> the simulator.
> 
> This sounds like a description of gnuplot. xmgrace would fit too. Both of 
> them are pretty easy to drive from an application written in c. Both are 
> can produce publication quality output. Both are well documented. When I 
> had a similar task, I went for xmgrace, because its GUI is active while 
> data from the application is streaming in. 

Here are some questions I have about any tool for waveform viewing:

Can you update plots as more data becomes available?

Is a cursor readout supported?  In my experience this has traditionally 
been a problem with matlab/octave/scilab.  If you can have a custom 
cursor function, you're way ahead of the commercial tools.  By custom 
function, I mean something which can read the data for a point on the 
graph and process through some function before displaying it.  Think 
reflection coefficients vs impedance on a smith chart, open vs closed 
loop response on a nichols chart, and others.

Custom gridlines?  (Rectangular, Polar, Nichols, Smith, others)

Is it easy to zoom in and out?

Can you scroll once you're zoomed in?  This seems to be lacking in 
matlab/octave/scilab but nearly all simulation waveform viewers have it.

Can large amounts of data be handled (thousands and thousands of points)?

Can it read out one or two waveforms out of a file that may have 
thousands?  I bring this up because it is quite easy with real world 
designs to end up with a multi-gigabyte data file.  If you have to load 
the entire thing to plot one waveform out of several thousand, you can 
take a real performance hit.

As Al mentioned, the ability to program functions is pretty darn 
important.  This is an area where tools like matlab/octave/scilab are in 
good shape and gwave (along with many other waveform viewers) is lacking.

> Any reason to reinvent the wheel?

We'd like a round one?  Don't get me wrong, I think there are a bunch of 
"sort of close" tools around, but I haven't come across one which didn't 
have some sort of major (to me at least) deficiency.   Even in the high 
cost arena, the cadence waveform viewer is almost good but has some 
areas where it stinks.

-Dan


_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user