- Chimera-users - RBVI Mailing Lists

Memory requirements for Chimera
by Tom Goddard 21 Feb '07

21 Feb '07

Hi Christopher, I think it starts making sense to use a 64-bit machine if you are loading data that requires >= 1 Gbyte. In general, 32-bit linux binaries will run on a 64-bit machine. Currently we only provide a 32-bit linux Chimera distribution though we plan on adding 64-bit Linux as a standard Chimera platform in the next 6 months. The 32-bit Chimera runs fine on 64-bit linux. The advantage of a 64-bit version is that it uses memory pointers that are 64 bits so they can address more than 4 Gbytes. On a 32-bit system the operating system and the many libraries that Chimera uses (~50) may occupy a significant portion of the 4 Gbyte address space. Even if you have less than 4 Gbytes of physical memory the operating system can and does use the full 32-bit 4 Gbyte address space and swap data to disk when there is not enough physical memory. But if a program asks for a 512 Mbyte block of memory and there is no available segment of that size in the address space then it will fail. Graphics drivers use the computer's main memory for data that will not fit on the card. Either 256 or 512 Mbytes on the graphics card should give good performance even for large systems. Unless you are using the volume "solid" rendering mode the rendering speed will slow to much less than 1 frame per second before you use up 256 Mbytes. For reference, the machines I currently use for developing Chimera and looking at large volume data sets are: 1) Mac dual G5 desktop with 1.5 Gbyte memory and 128 Mb ATI Radeon 9800 Pro graphics, and 2) MacBook Pro, Intel Core 2 duo, 2 Gbyte memory, 256 Mb ATI X1600 graphics. Thanks for the apoptosome papers -- interesting work. Tom Christopher Akey wrote: > Tom- > ok, I will keep notes on what is not working and send them to you, as > we go through making figures this week. > > Will chimera run fine on a 64 bit Linux box? The other option is 32 > bits, but if I buy a high end workstation it should probably be 64 bit > yes? > > Also, about making Chimera able to handle big sessions: is the rate > limiting factor the RAM or the memory on the Graphics Card. We have a > 512 Mb cache on our current graphics card and still have problems, so > I presume we are RAM limited. > > When buying a new machine, should I invest in a 512 Mb Nvidia card or > is 256 fine, as long as there is enough RAM? > > By the way, I am attaching two pdfs of papers published using Chimera > graphics. > > thanks. > > CWA > > > Tom Goddard wrote: >> Hi Christopher, >> >> I'm not sure what causes crashes in Chimera when you are working >> close to the available memory limit. If a memory allocation fails in >> C++ it should produce an error message in Chimera indicating that >> memory could not be allocated because we catch those exceptions. >> Memory allocations in C (not C++) libraries that Chimera uses are >> likely to cause a crash though. >> >> It would help us to know what Chimera is doing when the crashes >> occur. We may be able to put some safety checks in our C++ where C >> library allocations are known to fail. Do I understand correctly >> that crashes happen while you are saving an image (individual tiles >> being displayed on screen)? Chimera by default makes an image 3 >> times larger then requested to make a smoother image. That can take >> a good bit of memory. Are there other actions that cause crashes? >> >> Tom > >

1 0

An NMR Average Structure
by S W Simon Sham 21 Feb '07

21 Feb '07

Hi, I need to calculate an average structure from an emsembled structures in Chimera. Can we do it it Chimera? If yes, how? Thanks for any helps. Simon ------------------------------------------------------------------------ ----- Simon Sham, Ph.D. NMR Core Facility Comprehensive Cancer Center Univerisity of Alabama 933 South 19th Street CH19 B31 Birmingham, AL 35294 Office: (205) 934-5696 Fax : (205) 934-6475 Email : ssham44(a)uab.edu

2 1

Create surface from markers
by Tom Goddard 20 Feb '07

20 Feb '07

Hi Jeff, Chimera cannot currently create a surface from markers placed with the volume path tracer tool. It would not be so hard to write some Python code to do this. Here is an example of Python code that creates a surface in Chimera: http://www.cgl.ucsf.edu/chimera/docs/ProgrammersGuide/Reference/surface.html Tom Jeff wrote: > hi Tom, > > is there a way to use a marker file just to render a surface (a > series/group of dots, just needing to be connected by a mesh)? for > instance, to represent a membrane. > > -Jeff

1 0

Saving full map when zoning
by Tom Goddard 20 Feb '07

20 Feb '07

Hi Jean-Francois, You can save a map with the original map size while using volume zone by first pressing the the "Full" button in the Features / Subregion Selection panel of the volume dialog. Then use the File / Save Map As button. You can see what region will be saved by looking at the Features / Region Bounds panel (and you could type in new bounds there and press the Enter key instead of pressing the Full button). The volume zone capability (Features / Zone) normally uses a smaller box of data so that it is faster. Even when part of the surface is hidden by the zoning the whole surface within the box being used is computed and that is sometimes slow. Tom jean-francois menetret wrote: > > Hi Thomas, > > when using the zone option in Chimera, is it possible to have the size > of the saved mrc file (after zoning) be the same as the original file ? > > Best wishes > > Jean-Francois

1 0

Re: [Chimera-users] demo made under Linux has bland atoms in Windows
by Elaine Meng 17 Feb '07

17 Feb '07

On Feb 16, 2007, at 7:55 PM, Kenward Vaughan wrote: > For me, the byatom coloration is there to begin with under Linux. The > ach shows up colored properly, as does the acetylcholinesterase. > Aha!! Problem solved!! I am 99.99% sure the difference is your preferences file. The demo itself does not say to color by element when the structure is opened. However, there is a New Molecules section of the Preferences in which you can say structures should be colored that way when they are opened. Most likely you have this set on your home Linux machine, but not in the Chimera preferences used on the Windows machines. Months ago we discussed the need for some way to indicate that preferences should be bypassed at startup and/or when opening structures. This option would be appropriate for use with demos or scripts to be shared among multiple users each having some arbitrary set of preferences settings. However, this option hasn't been added (nor are we sure how best to implement it)... thanks for the additional motivation to revisit that issue! The demos included with Chimera were made in a directory with a default preferences file. It is likely, however, that some users with nondefault New Molecules preferences have played them back and gotten a result different than intended. > > I've noticed rotation is slow under Windows, but v. nice with Linux > (about 7 seconds total). > It depends on the graphics capabilities of the system. It is slow on our home desktop computer (>5 years old - you can almost hear gears grinding) but fast on my laptop (1-2 years old). Happy demoing and please do contact us with any problems or suggestions, Elaine ----- Elaine C. Meng, Ph.D. meng(a)cgl.ucsf.edu UCSF Computer Graphics Lab and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco http://www.cgl.ucsf.edu/home/meng/index.html

1 0

demo made under Linux has bland atoms in Windows
by Kenward Vaughan 16 Feb '07

16 Feb '07

...in a fashion, that is. A demo I plan on using at school (a Windows fortress, it is) displays all atoms as grey--no color (by element). The demo was created on a Linux machine (at my home). :( Any thoughts? Kenward ps. both versions of Chimera are the latest production version (2304). -- In a completely rational society, the best of us would aspire to be _teachers_ and the rest of us would have to settle for something less, because passing civilization along from one generation to the next ought to be the highest honor and the highest responsibility anyone could have. - Lee Iacocca

3 2

saving map in different position
by Gabor Papai 16 Feb '07

16 Feb '07

Hi Tom, Is there a way to save a map with only transformation, or obtain the transformation matrix if I do a freehand transformation of an object? We would like to create a composite volume of several spheres. It would be the best to write a single map of several differently positioned spheres but I could also apply the transformation in IMAGIC if I would know the matrix. Thanks for helping! Regards, Gabor -- Gabor Papai IGBMC Department of Structural Biology and Genomics 1, rue Laurent Fries, BP 10142 67404 Illkirch, France phone +33-3-90244796 Fax +33-3-88653201 E-mail: papai(a)igbmc.u-strasbg.fr

3 2

Re: [Chimera-users] making movies with MorphMap, exporting povray scenes, FitMapInMap
by Gatsogiannis, Christos 15 Feb '07

15 Feb '07

Hello Greg and Tom, povray: the assumed gamma-change was indeed the problem. Thank you very much! I´m waitin for the next snapshot release MorphMap: the updated version works correctly! Thanks! Do you maybe plan in the future on offering the possibility to save a density map also in different formats than mrc (situs, imagic .. etc) ? That would be very useful... Thanks again, Christos

2 1

cavity cropped view
by Randy Heiland 15 Feb '07

15 Feb '07

What options exist in Chimera for letting me show only that portion of a protein that is "near" another smaller molecule? Specifically, I have a ligand in a cavity of a large protein and would like to eliminate the "clutter" of the protein away from the cavity of interest. Thanks, Randy

3 8

making movies with MorphMap, exporting povray scenes, FitMapInMap
by Gatsogiannis, Christos 14 Feb '07

14 Feb '07

Dear Chimera users and developers, recently I used the new morph map tool and it worked very nice in my case. I encountered a problem while exporting the animation in any movie-format within Morphmap, I just get empty or very small files, which cannot be displayed in any media-player. I decided then to use the chimera-movie recorder utility to record the morphing animation and this worked fine (I´ve tested this with mpeg-2 format). At the moment I´m using the chimera-version 1.2309 in a SUSE LINUX 9.2 environment. Is this a version problem or a bug? Furthermore, I´ve exported some scenes in the povray format and I´ve noticed that the final images are of lower quality in comparison with images made some months ago with older chimera versions. I haven´t changed the povray-version, the rendering settings and the settings for the visualisation of my data in chimera (brightness etc...). The colors are too pale, I could send you some example-images if needed. I´ve tried also the Fit Map in Map tool recently and it worked perfectly (you sent this tool to a group-colleague some months ago). I´m wondering why you didn´t put this tool on the experimental site yet....and one more last question...after using Fit Map in Map, is there any possibility to save the "fitted" map with it´s new orientation? Thanks, Christos Christos Gatsogiannis Institute of Zoology Johannes Gutenberg University 55099 Mainz, Germany Tel: +49(0)6131-39-23091 email: gatsogia(a)uni-mainz.de

3 2