|
11/18/2008 09:31 AM
|
|
11/3/2008 02:23 PM
|
|
10/8/2008 06:21 PM
|
|
9/30/2008 08:17 PM
|
|
8/18/2008 07:48 PM
|
|
8/12/2008 02:30 PM
|
|
8/4/2008 11:56 PM
|
|
6/23/2008 12:11 AM
|
|
6/20/2008 08:48 AM
|
|
5/15/2008 05:27 AM
|
|
5/14/2008 06:13 PM
|
|
4/13/2008 03:19 PM
|
|
1/31/2008 06:12 PM
|
|
1/14/2008 09:43 PM
|
|
12/5/2007 07:23 PM
|
|
11/25/2007 06:48 PM
|
|
11/15/2007 05:30 AM
|
|
11/13/2007 07:28 AM
|
|
11/9/2007 07:29 AM
|
|
11/5/2007 04:49 PM
|
|
10/31/2007 07:06 PM
|
|
9/3/2007 05:04 AM
|
|
8/14/2007 04:35 PM
|
|
7/9/2007 05:24 PM
|
|
5/20/2007 07:13 PM
|
|
3/28/2007 03:00 AM
|
|
3/22/2007 05:25 PM
|
|
3/1/2007 09:26 PM
|
|
2/23/2007 07:09 PM
|
|
1/31/2007 07:05 PM
|
|
Pymol: HighResolution Raytracing on large Protein Complexes? [View Printable]
|
phhuber
Group: Member Posts: 6 Joined: Apr 24, 2007
|
Hi
I want to make a high resolution image of a large protein complex, its the capsid of parvovirus B19 consisting of 60 monomers (pdb code 1S58). I assembled the biological unit using split_states, calculated the surface and saved the png image using the png command, which works fine, but only generates the resolution which is shown on the screen with the OpenGL renderer. When I use the function ray 1000,1000, which should generate a 1000x1000 px image with the built-in raytrace renderer, Pymol shuts down with the error message:
Runtime Error! This application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information.
How could I solve this problem? I'm using Windows 2000 on a 2.8 GHz Pentium4 CPU, but had the same problem on WinXP with 2x3.0 GHz PentiumD. Actually the OpenGL-Quality would be sufficient to me, but not the screen resolution, would rather like a 5000x5000 px or more picture...And I don't mind if the computer has to work at night... ;-)
Thanks for your help! cheers phil
|
.........................
|
Posted Apr 24, 2007, 14:26 PM |
|
|
|
bgood
Group: Moderators Posts: 155 Joined: Apr 12, 2006
|
Hi Phil,
I'm not a Pymol user at the moment, so can't offer any specific advice :(. But, when things like this have happened to me before its almost always a result of running out of RAM for one reason or another. How much memory do you have available on your system? Can you check to see if its all being used?
I'm not a python user either.. can you specify the amount of memory available for the python interpretter? In java, you often have to increase the maximum heap size to run applications like this because, by default, the jvm is kept quite light (I believe 2mb).
|
.........................
|
| Posted Apr 24, 2007, 15:20 PM |
|
|
|
phhuber
Group: Member Posts: 6 Joined: Apr 24, 2007
|
Hi and thanks for the reply!
I have 1 GB of RAM and checking the used memory in the task manager reveals, that indeed Pymol uses a lot of RAM. Sometimes the available RAM drops to 5 MB, mostly it is about 20. This is also the case when I am rendering for instance the protein complex as sticks, which does not lead to a crash. When I render the complex as surface model, Pymol crashes IMMEDIATELY, I can't discover a complete RAM blockade in the task manager.
I unfortunately don't know how to check the maximum available memory for the python interpreter. Can you tell me how to do this?
cheers Phil
|
.........................
|
| Posted Apr 26, 2007, 14:24 PM |
|
|
|
bgood
Group: Moderators Posts: 155 Joined: Apr 12, 2006
|
Hi,
I think there is probably nothing you can do to python like what I was thinking before.. But, perhaps this from an old message on the pymol list will help?
"You may try to decrease the value of hash_max, which will reduce the memory requirement for ray tracing, at the expenso of longer rendering time. E.g:
set hash_max, 50 "
|
.........................
|
| Posted Apr 26, 2007, 14:37 PM |
|
|
|
ryan_m
Group: Moderators Posts: 284 Joined: May 06, 2006
|
If you were doing this on Linux you could use "nice" to limit the resources that Pymol uses. My guess is even without niceing the job you would not crash the system but rather just use swap memory until the job completes.
|
.........................
|
| Posted Apr 26, 2007, 17:00 PM |
|
|
|
|
|
frasermoss
Group: Admin Posts: 722 Joined: Feb 22, 2005
|
so for your 1000 x 1000 you should type in the GUI
png yourproteinname, dpi=1000
I have saved 1200 dpi images this way on my laptop running Windows XP, a 1.73GHz Pentium M with 1.5 GB RAM without hassle.
|
......................... "Opportunity is missed by most people because it is dressed in overalls and looks like work". Edison
|
| Posted Apr 26, 2007, 17:50 PM |
|
|
|
phhuber
Group: Member Posts: 6 Joined: Apr 24, 2007
|
| frasermoss said: | so for your 1000 x 1000 you should type in the GUI
png yourproteinname, dpi=1000
I have saved 1200 dpi images this way on my laptop running Windows XP, a 1.73GHz Pentium M with 1.5 GB RAM without hassle.
|
This is a nice command, but it does not alter the dimension of a picture, e.g. if the picture on your screen is 700x900px, the png command does not change that. It just makes the picture smaller with higher resolution or bigger with lower resolution. What I would like to do, is to really increase the number of pixels... so I can print it for instance as DIN A3 format.
|
.........................
|
| Posted May 01, 2007, 12:42 PM |
|
|
|
phhuber
Group: Member Posts: 6 Joined: Apr 24, 2007
|
| bgood said: | Hi, "You may try to decrease the value of hash_max, which will reduce the memory requirement for ray tracing, at the expenso of longer rendering time. E.g:
set hash_max, 50 "
|
I tried with the has_max set to 50 and 20, both of this didn't work as well... neither on linux.
|
.........................
|
| Posted May 01, 2007, 12:49 PM |
|
|
|
phhuber
Group: Member Posts: 6 Joined: Apr 24, 2007
|
| ryan_m said: | | If you were doing this on Linux you could use "nice" to limit the resources that Pymol uses. My guess is even without niceing the job you would not crash the system but rather just use swap memory until the job completes. |
I also tried to run the whole thing on Suse Linux 10.1 64bit. Pymol crashed as well, the application terminated without any comment... I have only very limited linux skills, but I will try to start pymol with increased "nice" increment and post my experience here... ;-) Thanks for all the good advice! Perhaps it's just a little too keen to render a whole virus capsid...? Should I wait for quantum computers? :)
|
.........................
|
| Posted May 01, 2007, 12:53 PM |
|
|
|
frasermoss
Group: Admin Posts: 722 Joined: Feb 22, 2005
|
| phhuber said: | | frasermoss said: | so for your 1000 x 1000 you should type in the GUI
png yourproteinname, dpi=1000
I have saved 1200 dpi images this way on my laptop running Windows XP, a 1.73GHz Pentium M with 1.5 GB RAM without hassle.
|
This is a nice command, but it does not alter the dimension of a picture, e.g. if the picture on your screen is 700x900px, the png command does not change that. It just makes the picture smaller with higher resolution or bigger with lower resolution. What I would like to do, is to really increase the number of pixels... so I can print it for instance as DIN A3 format. |
Can't you just use Photoshop once you've saved your high dpi resolution image to resize the file, maintaining the dpi? Use the Image >Image Size command and set desired image size and resolution and ask it to re sample. Or does that lose resolution for you?
|
......................... "Opportunity is missed by most people because it is dressed in overalls and looks like work". Edison
|
| Posted May 02, 2007, 16:22 PM |
|
|
|
phhuber
Group: Member Posts: 6 Joined: Apr 24, 2007
|
What I need to get a sharp AND large image, is a large pixel x pixel dimension. The resolution itself is not important, it can be, as you say altered in Photoshop. When you have a 1000x1000 px image, you can define it as 10x10 inch with a resolution of 100 ppi, or 1x1 inch with 1000 ppi... I need a picture in the size of A3 (approx. 30x42 cm) with a printable resolution, let's say 300 dpi. When I now just enlarge a small image with Photoshop, the quality does of course not increase, only the size (e.g. the resolution decreases with increasing size...) Well, I'll anyhow try Photoshop's resample function, as this tool achieves amazing results sometimes :)
|
.........................
|
| Posted May 03, 2007, 9:34 AM |
|
|
|
|
top of page
|
|
Forum Jump
| |