I'm saying that the plots generated from my local J7 are noticeably slower
than the J6 equivalent.  I just now timed (after loading "plot"):

   'type surface' plot +/~2 o. i:5j99

In J6 (console mode running jee), this is nearly instantaneous - perhaps 0.5
s;  in J7, it takes about 3 s.  Similarly, resizing the plot
window is virtually immediate in J6 but takes about 2 s in J7.  I also
notice that J6 remembers the last size to which I resized the plot window -
a subsequent plot after I've killed the plot window comes up in the same
place at the same size to which the previous plot was set - whereas J7
brings up the same size plot each time.  In the past, I've found this (J6)
persistence useful.

I was assuming that this is due to the change to GTK but don't really know
why.  I haven't looked into the nitty-gritty of the GTK stuff yet but this
may motivate me to do so.  I see by burrowing down into the "jwplot"
namespace that the high-level code I've perused looks the same between the
two versions, so this is encouraging.  However, I also see that there are
obvious differences as
   (#@:nl_jwplot_) &> 0 1 2 3
returns "325 0 0 9" in J6 but "173 0 0 0" in J7.

On Thu, Feb 10, 2011 at 11:52 AM, Eric Iverson <[email protected]>wrote:

> Are you saying that jhs GC plots are much slower in the last few weeks over
> earlier versions?
>
> This surprises me. However, there was a change that could be investigated.
> Some weeks ago jhs GC plot was changed to use a more compact encoding of
> data sent to the GC server. This allowed more data (2 to 3 times) to be
> sent
> to the chart server. The encoding is pretty simple and I doubt it makes any
> timing difference at all on the J side. I guess there is a small chance the
> GC server could be slower in doing the decoding. All in all seems unlikely.
> I suspect network delays or perhaps a loading problem at the GC server.
>
> On Thu, Feb 10, 2011 at 10:46 AM, Devon McCormick <[email protected]
> >wrote:
>
> > I mentioned at NYCJUG the other night that I've noticed slowness in
> > plotting
> > compared to the older version - both in the initial rendering and when
> you
> > manually resize the window.
> >
> > On Thu, Feb 10, 2011 at 10:18 AM, Daniel Durieu <[email protected]
> > >wrote:
> >
> > > -Lab 1 : A J introduction
> > >  is working perfectly under JHS
> > >  just one remark: from time to time plotting is very slow
> > > ...
> > >
> >
> >
> > --
> > Devon McCormick, CFA
> > ^me^ at acm.
> > org is my
> > preferred e-mail
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>



-- 
Devon McCormick, CFA
^me^ at acm.
org is my
preferred e-mail
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to