On Mon, 3 Mar 2008 13:06:45 -0600 "Paul Johnson" <[EMAIL PROTECTED]> babbled:
11:40AM ~/.e/e/backgrounds > pwd /home/raster/.e/e/backgrounds 11:40AM ~/.e/e/backgrounds > /bin/ls *.edj | wc -l 100 11:40AM ~/.e/e/backgrounds > for I in *.edj; do enlightenment_remote -default-bg-set $HOME/.e/e/backgrounds/$I; sleep 1; done; enlightenment_remote -default-bg-set "" [... wait for it to set all 100 and go back to theme default...] i'm running e under memprof. 0 memleak. at the end e settles back to exatly the memory it had allocated before. its VSIZE does expand as it often needs a lot of ram (16m+) to store bg's (a 1600x1200 wallpaper uses about 8mb of ram. u will find u have 2 of such beasties in ram when changing backgrounds and 1 will hang around 1 render cycle until it is cleared). if you want the details. here is e before i start this 100 bg flip (from top and memprof) and after: --before-- top: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 22723 raster 15 0 52604 10m 6328 S 0 1.1 0:08.83 enlightenment memprof: # of Allocations: 32375 Total Bytes: 3332570 --after-- top" PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 22723 raster 15 0 52520 10m 6348 S 0 1.1 3:55.53 enlightenment memprof: # of Allocations: 32431 Total Bytes: 3337863 sure... a few more allocations - likely for caching and initial one-demand loading of a few extra bits of data, but after a mammoth run of changing 100 wallpapers (a lot of them are 1600x1200 or 1920x1200 images others are full edje designs with multiple images, layers etc., and a few are even animated), e has used (according to memprof) an extra 5293 bytes. given that there are 10000's of dynamic allocations during this run, the fact that RES (resident memory - actual pages of memory really read from or written to) is still at 10m (ok - we dont see it in more detail- but it's not gone anywhere significant) and 6.3m of that memory is shared (glibc for example, libjpeg, libX11, libpng, libevas etc. of which that memory is split and used again for all processes using that library, so generally such costs should be discounted from an actual memory profile and more considered as part of a base system cost). so... you have something wrong. this is e from cvs. i built it and updated just before i did this run. it doesn't get much fresher than that. i use no moduels that didn't ship with e itself (e_modules do not ship with e - they are 3rd party). so personally, as a developer - i'm happy with e there - it's leak-free in this department, and that makes me all warm and fuzzy :) > Hello again, E users: > > I had an interesting E session yesterday. I was stress testing a > script I wrote to create edj files and put them on workspace > backgrounds. I was changing backgrounds frequently, and I noticed the > system was getting sluggish, but didn't realize the problem until I > tried to use an E menu to start a terminal and a popup from E said it > could not fork a process. Hm. In an existing xterm I ran top and > found that E was using 85% of my system's 1 gig of memory and 2 gig of > swap was full. > > I restarted several times and re-tested and E's memory consumption > would begin around 1.5% of the system, which is fine. After flushing > through several background changes, it appeared to me that E's memory > consumption grows about 3% each time the backgrounds are changed (on a > system with 8 desktops). > > I'm attaching the script I was using and the edc file it requires to > make edj files. I wonder if you run this over and over, > > epjbackground.pl -c Y -w 2 -d /usr/local/share/Backgrounds > epjbackground.pl -c Y -w 3 -d /usr/local/share/Backgrounds > epjbackground.pl -c Y -w 4 -d /usr/local/share/Backgrounds > [and so forth for all workspaces] > > > do you see the same thing? > > This happens in a Fedora 8 system running > > enlightenment-startup-1.0-2.fc8.prof_k > enlightenment-backgrounds-1.0-2.fc8.prof_k > enlightenment-0.16.999.042-6.20080225cvs.fc8.prof_k > > How to diagnose? Does enlightenment_remote have a way to report on memory > usage? > > I tested this running E as a stand-alone desktop environment as well > as a window manager within a Gnome desktop. Same outcome. > > > PJ > > -- > Paul E. Johnson > Professor, Political Science > 1541 Lilac Lane, Room 504 > University of Kansas > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) [EMAIL PROTECTED] ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ enlightenment-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/enlightenment-users
