I've run into similar situations in the past, although in my case usually due to large maps (a 3.5 MB pdb file, or even a set of them, typically requires less memory than a ~50MB electron density map). In my experience, it's much better to load things from scripts (pml) rather than storing things in session files (pse). I may be remembering incorrectly, but from previous mailing list discussions I believe the higher memory usage for pse files was due to overhead for python de-pickling.
You may also want to check your graphics drivers - a friend of mine was having some pymol performance problems that turned out to be driver related. Pete John Ladasky wrote: > I'm currently using PyMol 1.4.1 on Ubuntu Linux 11.10. My hardware is fairly > up-to-date, an AMD 6-core, 3.2 GHz CPU ,and 4 GB of DDR1200 RAM. My only old > hardware is my NVidia 8800 GTS graphics card. I have started to assemble > rather large visualizations. I am noticing CPU performance lags, and > unexpectedly high memory usage. In my hands, PyMol is behaving like the > infamous Flash Player browser plug-in -- hogging memory and CPU cycles (as > well as disk space) for no obvious reason. > > Here's an example. I have superimposed 15 PDB files which are snapshots from > GROMACS molecular dynamics simulations. These contain protein molecules in a > water box. I admit, the PDB files are pretty large, around 3.5 MB each. > Still, 15 x 3.5 MB = 52.5 MB. That should be manageable, I think. > Ambitious, but still within reasonable limits. > > But what is the size of the PyMol session (.pse) file that I built from this > stack of PDB files? 234 MB! Why did this bloat up to fully five times the > size of the source data (which is already in plain text, and full of white > space)? > > What's more, when I load this 234-MB file into PyMol, the Python interpreter > which PyMol invokes grabs a full 1.0 GB of memory. The CPU usage of that > Python interpreter (single-core, of course) immediately jumps to around 40% > even when I'm not asking PyMol to do anything. After I've worked with PyMol > a bit with this file loaded, CPU usage will jump to 100%, and it stays there. > Once that happens, the cursor will lag, jitter or freeze, and I can end up > selecting unintended menu items. Keep in mind, this is with a completely > stationary image in the PyMol Viewer window. > > I'm looking to reduce all of this overhead. Excluding some water molecules > from the source PDB files comes to mind. Perhaps I can find a way to clip my > simulation box (I would like to retain the option to visualize any water > molecules which actually contact the protein). However, I seem to have more > fundamental problems than excess water molecules. Alternately, if there are > settings in PyMol that I can adjust to get it to behave, that would be great. > > Thanks for any advice you may have! > ------------------------------------------------------------------------------ Virtualization & Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/ _______________________________________________ PyMOL-users mailing list (PyMOL-users@lists.sourceforge.net) Info Page: https://lists.sourceforge.net/lists/listinfo/pymol-users Archives: http://www.mail-archive.com/pymol-users@lists.sourceforge.net
