Hello Nikolaos,

> I have a question: which waveform format in your opinion, is either  
> sufficient or could be extended to support a representation for analog  
> data (real arithmetic)? It can even be done via bitstrings of course,  
> but then a specifier would be needed (to denote the use of analog  
> data). There are also some other possible extensions that come in mind.
> 
> I agree that (E)VCD is not ideal for VHDL.

As far as I understand what you are looking for, in a technical point of
view only, you want a waveform format that is able to contain
floating-point data values.
You also want this format to be handled by a waveform viewer, say
GTKWave, so that you can use it out of the box.

Am I right here ?

GTKWave already handles real values from the VCD format. Note that I
don't remember whether it's standard or not. At least the comments in
the GTKWave source code suggest it is.
To generate VCD files with floating-point data values, use the "r" or
"R" token, it is read as a double by GTKWave (in v3.2.3, file vcd.h line
955). Example of syntax:

#100
r42.123 &1

This works like charm.

Best regards,
Adrien Prost-Boucle



_______________________________________________
Ghdl-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/ghdl-discuss

Reply via email to