Does Chimera have a fixed amount of memory allocated to it, and if so, is it possible to increase this amount? I am working with very large system sizes and am trying to reduce the amount of time it takes to perform certain operations.
Chimera does not have any fixed amount of memory allocated. It will use whatever is available. If you load large molecular systems (100,000 atoms) or large volume data sets (> 100 Mbytes) your computer may start swapping (transfering data back and forth to disk because it does not fit in memory). That can dramatically slow down any computation by factors of 10 or 100. The only real solution is to use a machine with more memory, or load smaller systems in Chimera. Are you working with molecules or volume data? There are tricks to reducing the data size depending on the type of data. Tom
I also encounter memory problems (quite often) because I'm working with large structures such as the ribosome. When I try to restore .py sessions as large as 40 Mb, Chimera terminates after around 5-10 minutes and throws a memory error. It doesn't help that my machine runs on dual pentium 4 processors (2.8 GHz each) and 2.0 Gb RAM. I just gave up saving too large sessions in my PC when working with whole ribosome structures. I have had some correspondence a long time ago with Eric Pettersen regarding this. He suggested me the following: "....if you have access to another Windows machine with more that 1 GB of memory, you might try restoring the .py session file there and see if it works and produces a .pyc file. If it does, you could copy the .pyc file back to your machine and be able to restore the session while bypassing the compilation step. Another alternative if you are a little Python-savvy is to open Chimera's Python shell (Tools->Programming->IDLE) and compile the .pyc file yourself with: import py_compile py_compile.compile('full path to .py file', 'same path except with .pyc') [remembering that backslashes in quoted strings need to be doubled] then stop Chimera and start a new one in which you try to restore the session. The important difference between .py and .pyc files is that .py files are portable between different computer architectures and therefore can safely be distributed to other people and/or used on other machine types whereas .pyc files are not guaranteed to be portable." I have not really tried the trick above since at the moment, I'm the only one here at the lab with a PC with 2 GB RAM...and still I couldn't restore my large sessions. I have been trying to install Chimera in our linux cluster here but have had some setbacks with missing libraries,,,and some other things to do... :( ./jonathan ***************************************** Jonathan LS Esguerra, MSc Bioinformatics PhD student, Microbiology/Functional Genomics Cell and Molecular Biology-Microbiology Göteborgs University Box 462, Medicinaregatan 9C 40530 Gothenburg City Sweden Tel:+46 (0)31-7733738 Fax:+46 (0)31-7732599 ***************************************** ----- Original Message ----- From: "Thomas Goddard" <goddard@cgl.ucsf.edu> To: <brmorgan@clarku.edu> Cc: <chimera-users@cgl.ucsf.edu> Sent: Tuesday, September 12, 2006 10:39 PM Subject: Re: [Chimera-users] Memory allocation
Chimera does not have any fixed amount of memory allocated. It will use whatever is available. If you load large molecular systems (100,000 atoms) or large volume data sets (> 100 Mbytes) your computer may start swapping (transfering data back and forth to disk because it does not fit in memory). That can dramatically slow down any computation by factors of 10 or 100. The only real solution is to use a machine with more memory, or load smaller systems in Chimera.
Are you working with molecules or volume data? There are tricks to reducing the data size depending on the type of data.
Tom
Hi Jonathan, I guess when I put out the announcement for the 1.2255 snapshot release I forgot to mention that the session code got rewritten, so you should try that release. Using 1.2255, sessions are about 40% smaller, use 15% less memory and take half the time to start up. Unfortunately, Brittany's problem is that Chimera isn't very efficient when handling many thousands of small molecules (compared to a large molecule with about the same number of atoms as the small molecules combined). We're looking into that. --Eric On Sep 12, 2006, at 4:08 PM, Jonathan LS Esguerra wrote:
I also encounter memory problems (quite often) because I'm working with large structures such as the ribosome. When I try to restore .py sessions as large as 40 Mb, Chimera terminates after around 5-10 minutes and throws a memory error. It doesn't help that my machine runs on dual pentium 4 processors (2.8 GHz each) and 2.0 Gb RAM. I just gave up saving too large sessions in my PC when working with whole ribosome structures.
I have had some correspondence a long time ago with Eric Pettersen regarding this. He suggested me the following:
"....if you have access to another Windows machine with more that 1 GB of memory, you might try restoring the .py session file there and see if it works and produces a .pyc file. If it does, you could copy the .pyc file back to your machine and be able to restore the session while bypassing the compilation step.
Another alternative if you are a little Python-savvy is to open Chimera's Python shell (Tools->Programming->IDLE) and compile the .pyc file yourself with:
import py_compile py_compile.compile('full path to .py file', 'same path except with .pyc') [remembering that backslashes in quoted strings need to be doubled] then stop Chimera and start a new one in which you try to restore the session.
The important difference between .py and .pyc files is that .py files are portable between different computer architectures and therefore can safely be distributed to other people and/or used on other machine types whereas .pyc files are not guaranteed to be portable."
I have not really tried the trick above since at the moment, I'm the only one here at the lab with a PC with 2 GB RAM...and still I couldn't restore my large sessions. I have been trying to install Chimera in our linux cluster here but have had some setbacks with missing libraries,,,and some other things to do... :(
./jonathan
***************************************** Jonathan LS Esguerra, MSc Bioinformatics PhD student, Microbiology/Functional Genomics Cell and Molecular Biology-Microbiology Göteborgs University
Box 462, Medicinaregatan 9C 40530 Gothenburg City Sweden
Tel:+46 (0)31-7733738 Fax:+46 (0)31-7732599 *****************************************
----- Original Message ----- From: "Thomas Goddard" <goddard@cgl.ucsf.edu> To: <brmorgan@clarku.edu> Cc: <chimera-users@cgl.ucsf.edu> Sent: Tuesday, September 12, 2006 10:39 PM Subject: Re: [Chimera-users] Memory allocation
Chimera does not have any fixed amount of memory allocated. It will use whatever is available. If you load large molecular systems (100,000 atoms) or large volume data sets (> 100 Mbytes) your computer may start swapping (transfering data back and forth to disk because it does not fit in memory). That can dramatically slow down any computation by factors of 10 or 100. The only real solution is to use a machine with more memory, or load smaller systems in Chimera.
Are you working with molecules or volume data? There are tricks to reducing the data size depending on the type of data.
Tom
_______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://www.cgl.ucsf.edu/mailman/listinfo/chimera-users
participants (4)
-
brmorgan＠clarku.edu -
Eric Pettersen -
Jonathan LS Esguerra -
Thomas Goddard