[PyMOL] Does (PDB) size really matter? (Jason Vertrees)
Hi Warren and Jason. I've playing around once with nice movies but memory has been a problem. Now I got a dual Athlon with 2Mb Ram, but didn't test it yet, I'm afraid (such server cannot stop :-). Warren, you really have a great software and I really want to do with it what I've done with VMD. See an example: http://www.lac.inpe.br/~alan/IntroducingMD.avi I'm wondering if (and that's my suggestion to you Jason) it'd be a nice idea parting you movie in several takes (as I did for the example above), like playing-as-a-director-in-a-real-movie. When I get a spare time I'll try to get back to my movies. Cheers, Alan On Sun, 25 Apr 2004 pymol-users-requ...@lists.sourceforge.net wrote: Today's Topics: 1. Does (PDB) size really matter? (Jason Vertrees) Warren, Thanks! I was setting 'hash_max' higher and higher (but my memory monitor always showed at least 90 megs free at the least). I'll try the lower hash_max; I'll try lower this time. I think we have a really great movie in the works, and as I play with PyMol more and more, I can make much cooler things. You've got some great software here! My 0.96b2 PSE file is bzipped, for all to download, at http://www.vertrees.org/~tree/scene1.pse.bz2 it's about 2.7 MB; The 0.95 PSE file is http://www.vertrees.org/~tree/095_scene1.pse.bz2. atom_count is about 85,000 for them. The movie script is as in the previous message. As of now, I'll try smaller hash_max's and also if that's not good, reduce the size of my circular lipid bilayer. Many thanks for the quick response! Regards, -- Jason -- -- Alan Wilter Sousa da Silva -- D.Sc. - IBCCF/UFRJ Bolsista Pesquisador LAC-INPE São José dos Campos (SP), Brasil www.lac.inpe.br/~alan
RE: [PyMOL] Does (PDB) size really matter?
Hi Jason, I see you have hash_max set to 400 -- that's the problem. In empirical testing, I found that there is rarely a reason to go above 200 with this setting. Having a hash table that is too finely sampled is no better than having one too coarsely sampled. PyMOL's default of 100 is intended to avoid running out of RAM on systems with just 256 MB. If you have more RAM, then going up to 150 or 170 is usually worth it. 250 may sometimes be helpful, but 400 is almost certainly overkill. Because this is a 3D hash table, each 2X increase can potentially involve an 8X increase in memory required. set hash_max,200 set antialias,1 produces the following image in 204 seconds using 948 MB of RAM on a dual G5 http://delsci.com/img/scene1.jpg with set light, [-0.4,0.4,-1]: http://delsci.com/img/scene1_flip_light.jpg You can mellow out that techno-green and give your scene a more natural feel with: set ray_blend_green, 0.4 set ray_blend_colors, on http://delsci.com/img/scene1_mellow.jpg And then, it helps to back off on the green color itself, so that those spheres don't look so over-lit: set_color green, [0.25,0.75,0.25] rebuild Now you've got a reasonably balanced image... http://delsci.com/img/scene1_final.jpg Cheers, Warren -Original Message- From: pymol-users-ad...@lists.sourceforge.net [mailto:pymol-users-ad...@lists.sourceforge.net] On Behalf Of Jason Vertrees Sent: Saturday, April 24, 2004 11:03 PM To: PyMol Users List Subject: [PyMOL] Does (PDB) size really matter? Warren, Thanks! I was setting 'hash_max' higher and higher (but my memory monitor always showed at least 90 megs free at the least). I'll try the lower hash_max; I'll try lower this time. I think we have a really great movie in the works, and as I play with PyMol more and more, I can make much cooler things. You've got some great software here! My 0.96b2 PSE file is bzipped, for all to download, at http://www.vertrees.org/~tree/scene1.pse.bz2 it's about 2.7 MB; The 0.95 PSE file is http://www.vertrees.org/~tree/095_scene1.pse.bz2. atom_count is about 85,000 for them. The movie script is as in the previous message. As of now, I'll try smaller hash_max's and also if that's not good, reduce the size of my circular lipid bilayer. Many thanks for the quick response! Regards, -- Jason -- Jason Vertrees BSCB Graduate Student @ UTMB, Galveston javer...@utmb.edu :: http://www.bscb.utmb.edu --- This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek For a limited time only, get FREE Ground shipping on all orders of $35 or more. Hurry up and shop folks, this offer expires April 30th! http://www.thinkgeek.com/freeshipping/?cpg=12297 ___ PyMOL-users mailing list PyMOL-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pymol-users
[PyMOL] Does (PDB) size really matter?
Warren, Thanks! I was setting 'hash_max' higher and higher (but my memory monitor always showed at least 90 megs free at the least). I'll try the lower hash_max; I'll try lower this time. I think we have a really great movie in the works, and as I play with PyMol more and more, I can make much cooler things. You've got some great software here! My 0.96b2 PSE file is bzipped, for all to download, at http://www.vertrees.org/~tree/scene1.pse.bz2 it's about 2.7 MB; The 0.95 PSE file is http://www.vertrees.org/~tree/095_scene1.pse.bz2. atom_count is about 85,000 for them. The movie script is as in the previous message. As of now, I'll try smaller hash_max's and also if that's not good, reduce the size of my circular lipid bilayer. Many thanks for the quick response! Regards, -- Jason -- Jason Vertrees BSCB Graduate Student @ UTMB, Galveston javer...@utmb.edu :: http://www.bscb.utmb.edu
[PyMOL] Does (PDB) size really matter? (Help!)
PyMolers, Our setup is a Dual Xeon 3.2 GHz 1GB RAM and Nvidia Quadro running PyMol 0.95. (Also available, Dual Opteron 240 64-Bit Linux w/0.95 w/1 GB RAM w/same problems.) We are trying to render a nifty movie but can't. PyMol keeps crashing during the render process. The scene had originally 185,000 atoms, but now I've snipped out some cutting it down to 85,000 or so. The command ray 800, 600 dies. I then collected all of my bilayer segment objects into one larger one (so, instead of 8 objects I had one 8x as large, hoping that the reduction in the number of objects freed up some RAM.) The movie script for scene1 is http://www.vertrees.org/~tree/scene1.py. I'm very new to making movies, but it does what I want. Here's some PyMol output: PyMOLcount_atoms count_atoms: 86220 atoms PyMOLray VLAMalloc-ERR: realloc failed *** EEK! PyMOL just ran out of memory and crashed. To get around this, *** *** you may need to reduce the quality, size, or complexity of the scene *** *** that you are viewing or rendering.Sorry for the inconvenience... *** /usr/bin/pymol.com: line 2: 11531 Aborted /usr/bin/python /usr/lib64/python2.3/site-packages/pymol/__init__.py $* Any ideas on how we can keep the movie w/o sacrificing quality or number of atoms? If need be, I can probably trim another 3,000-10,000 (max) atoms out. Much TIA, -- Jason -- Jason Vertrees BSCB Graduate Student @ UTMB, Galveston javer...@utmb.edu :: http://www.bscb.utmb.edu
RE: [PyMOL] Does (PDB) size really matter? (Help!)
Ack! Sorry to hear this. The two things that determine memory usage when rendering are: 1) the total number of primitives (spheres, triangles, cylinder [i.e. lines] etc.) 2) the value of hash_max #1 may not be easily remedied for your scene #2 can be improved by setting hash_max to something like 65 or 50. Rendering will take much longer, but less RAM will be used. ~200k atoms really shouldn't be a problem on a machine with that kind of RAM, so there is also a chance here you might be experiencing a bug. It might be worth emailing me a (compressed?) copy of the session file just to be sure... I don't see anything wrong with your script. Cheers, Warren -Original Message- From: pymol-users-ad...@lists.sourceforge.net [mailto:pymol-users-ad...@lists.sourceforge.net] On Behalf Of Jason Vertrees Sent: Saturday, April 24, 2004 6:50 PM To: PyMol Users List Cc: Tzintzuni Garcia I.; Roger B. Sutton Subject: [PyMOL] Does (PDB) size really matter? (Help!) PyMolers, Our setup is a Dual Xeon 3.2 GHz 1GB RAM and Nvidia Quadro running PyMol 0.95. (Also available, Dual Opteron 240 64-Bit Linux w/0.95 w/1 GB RAM w/same problems.) We are trying to render a nifty movie but can't. PyMol keeps crashing during the render process. The scene had originally 185,000 atoms, but now I've snipped out some cutting it down to 85,000 or so. The command ray 800, 600 dies. I then collected all of my bilayer segment objects into one larger one (so, instead of 8 objects I had one 8x as large, hoping that the reduction in the number of objects freed up some RAM.) The movie script for scene1 is http://www.vertrees.org/~tree/scene1.py. I'm very new to making movies, but it does what I want. Here's some PyMol output: PyMOLcount_atoms count_atoms: 86220 atoms PyMOLray VLAMalloc-ERR: realloc failed ** ** *** EEK! PyMOL just ran out of memory and crashed. To get around this, *** *** you may need to reduce the quality, size, or complexity of the scene *** *** that you are viewing or rendering.Sorry for the inconvenience... *** ** ** /usr/bin/pymol.com: line 2: 11531 Aborted /usr/bin/python /usr/lib64/python2.3/site-packages/pymol/__init__.py $* Any ideas on how we can keep the movie w/o sacrificing quality or number of atoms? If need be, I can probably trim another 3,000-10,000 (max) atoms out. Much TIA, -- Jason -- Jason Vertrees BSCB Graduate Student @ UTMB, Galveston javer...@utmb.edu :: http://www.bscb.utmb.edu --- This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek For a limited time only, get FREE Ground shipping on all orders of $35 or more. Hurry up and shop folks, this offer expires April 30th! http://www.thinkgeek.com/freeshipping/?cpg=12297 ___ PyMOL-users mailing list PyMOL-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pymol-users
