On Tue, 3 Jun 2008, Timothy Normand Miller wrote:

> Let me begin by apologizing for being a moron and not making sure I
> had the latest version.  I didn't take my vitamins today.  :)  Also,
> Fink is _horribly_ out of date on this.

Fink's at 2.0.0pre5?  That's four years old.  I'm impressed.  =)  The
viewer is a bit different from what it used to be.


> On Tue, Jun 3, 2008 at 3:40 PM, Stephen Williams <[EMAIL PROTECTED]> wrote:
>
> > GTKwave is handy in that the LXT format is far more compact then
> > VCD, and thus works well with long/large simulations. Also, it
> > handles real values in nets, that barren VCD does not support.
> > But the, Windows portability is a bit of an issue at times. There
> > may be pre-built Windows binaries around if that is your issue.
>
> Ooh.  I'm going to try this LXT format!

FYI, in the 3.1 series, loading VCD files doesn't nearly take up as much
memory as it used to as it now uses a recoding loader.  For very large
files, you will be issued a warning and be told about the --giga option
which enables additional realtime memory compression.  That said, LXT
isn't always necessary for faster performance.  It really depends on the
input file.

Note also that if you have "ridiculously huge" VCD traces, you can try
this to convert to VZT using the VZT compressor:

vcd2vzt x.vcd y.vzt -d 9 -m 256 -z 1 -r

Then: gtkwave y.vzt --cpu=4
to enable the block-based SMP VZT loader for, say 4 cpus.  Tweak that
number as appropriate based on CPUs and/or physical memory size.

Also, --start and --end are handy for cutting down on memory usage if you
only need to view a subset of time in a trace, etc.

-t



_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev

Reply via email to