[IPython-user] 96% cpu usage with new scipy

Fernando Perez Fernando.Perez at colorado.edu
Thu Dec 8 11:26:19 CST 2005


Ryan Krauss wrote:
> I am running IPython 0.6.16.svn with new scipy.  As soon as I start
> Ipython, it starts using 96% of the cpu before I type any commands and
> it stays that high.  This happens only when I use the -pylab flag, and
> it happens whether I set numerix: Numeric, numarray, or scipy.  I am
> running matplotlib 0.85.1.cvs and scipy 0.4.3.1478 (new scipy).  It
> seems like this is a threading related issue, because if I start
> ipython without the -pylab flag and then import pylab and plot
> something, my cpu isn't at 96%, but I don't have the lovely ipython
> plot threading and have to close all my plots before I can do anything
> else.  I have no idea if this is a problem with my system,  ipython,
> or matplotlib.  (I don't think it is a new scipy problem, because I
> have this same behavior with the other numerix options).  Let me know
> if I can get any more information to help track this down.

Mmh, the first thing is to track down which GUI toolkits cause the problem. 
Try cycling your matplotlibrc backend choice amongst TkAgg, GTKAgg, WXAgg and 
QtAgg, and see which ones show the bad behavior.  That will start to provide 
useful info.

Cheers,

f




More information about the IPython-user mailing list