FW: [ccp4bb]: linux memory usage

[gwebster@caregroup.harvard.edu]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***



-----Original Message-----
From: Webster,Gordon (Pine/Pop User) 
Sent: Tuesday, June 19, 2001 4:30 PM
To: 'Kevin Battaile'
Subject: RE: [ccp4bb]: linux memory usage


Dear Kevin

I think that this problem was fixed in RH 7.1. I heard that the memory leak
originated in the code for the new gdk-pixbuf module (version 0.7.0) of the
Gnome desktop panel.

You might want to check the version numbers for the gdk sources on your
machine or try running RedHat's "memprof" utility that profiles memory usage
and detects leaks (you can get it from their website www.redhat.com).

Upgrading to RH 7.1 or downloading and recompiling later gdk sources
(version > 0.7.0) should fix it.

Gordon


-----Original Message-----
From: Kevin Battaile [mailto:battaile@post.its.mcw.edu]
Sent: Tuesday, June 19, 2001 3:11 PM
To: CNS board; CCP4 board
Subject: [ccp4bb]: linux memory usage


***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***


Hi,

I am experiencing an annoying problem with my linux box that is
rendering it almost unusable. If I start from freshly booted machine and
start submitting cns or ccp4 jobs, the memory slowly gets eaten up to the
point where the machine crashes. If I push the reset button to
restart, the machine boots up still indicating by 'top' that memory
is used up. Then if I do a regular reboot, it looks like I get a clean
slate. The machine doesn't indicate that it is swapping to disk.

Prior to my last reboot ( with only two x-terminals running) according to
'top' 95% of the memory is used, but at the same time the running list
indicated that X is the largest user of memory at only 23%. Here are some
of the numbers from top:

257600 av,  245248 used, 12356 free, 136376 buff, 59028 cache 

Then after reboot:

257600 av, 89188 used, 168412 free, 6269 buff, 45428 cache

FWIW, I am running RedHat 7.0 with a 1GHz AMD Athlon and 256 M RAM.

So .... does anyone have any ideas? This was an annoyance when I was
running cns scripts, but now that I have ccp4 working, I can run through
molrep once before I have to reboot, if I am lucky. 

Thanks,

Kevin
battaile@mcw.edu