[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Memory leak under Linux???

From: John Utz
Subject: Re: Memory leak under Linux???
Date: Sat, 27 Jan 1996 18:02:20 -0800 (PST)


On Fri, 26 Jan 1996, Dan Eldred wrote:

> I am encountering a pretty severe memory leak under Linux;
> each time I rerun a program that makes plots, Octave grabs
> another 1 meg of memory until all memory is used up.  The
> version of Gnuplot is 3.5 patchlevel

        hmm u mean the gnuplot is not stock 3.5? are u running a 
multiplot patch?

        I am running FreeBSD 2.0.5-RELEASE, and some indeterminate level 
of gnuplot, post 3.5, pre 3.6-beta. I plan to upgrade to the latest 
gnuplot 3.6-beta tonight or tomorrow...

  The Linux
> kernel is 2.1.13 (ELF).  Before I kill time investigating this
> further, I have to ask the group if anyone else has encountered
> this problem.

        Octave has been dieing after a couple of plots for me too, but i 
had attributed it to the fact that the work i was doing on it had just 
bumped up an order of magnitude in complexity.

> BTW I ran the same m-file on a Sun running SunOS and Octave 1.1.1,
> and didn't encounter the memory leak.

        This is a really tuff call. I hate doing this me too stuff, but i 
think it is incumbent upon me to pipe up, given the similarity of the 

        Obviously, now that we have agreed that we might have a problem, 
and we both share the X86 architecture, it is probably our job to fix it! :-)
Or, for that matter, to decide that the problem is that we are both too 
poor to stick suffficient ram on our machines, and we are just running 
out of swap :-(

        I myself am running a 386/33, 387, and 10 meg of ram, which is 
basically, chickenfeed.

        What do u have, and what are u doing? Could u post the m-file? 
Could i also suggest that u and i both get the latest gnuplot and see if 
that makes it go away?

        Finally, what in the heck are we supposed to do to fix it 
anyway? Is this something that mprof is supposed to help us track down?

        Does anybody in the list have any ideas where we would start?

> > Thanks, Dan    address@hidden > 

 John Utz       address@hidden
        idiocy is the impulse function in the convolution of life

reply via email to

[Prev in Thread] Current Thread [Next in Thread]