- Chimera-users - RBVI Mailing Lists

Mac laptop benchmark
by Tom Goddard 07 Feb '07

07 Feb '07

Hi Jeff, I ran the Chimera benchmark tool on my new MacBook Pro and got scores: surface 543 mesh 545 contour 183 solid 256 recolor 132 But running hand tests I see the real surface/mesh scores should be 380. It appears the benchmark no longer is reporting correct scores either because a graphics driver optimization is speeding things up because the test image is not actually displayed, or Chimera code has changed (perhaps no longer guaranteeing the OpenGL buffers are swapped). Probably other scores on the Chimera benchmark web page are likewise inaccurate. This will have to be investigated. At any rate, actual surface rendering is about 65% faster than my Dell Precision M50. A test on a GeForce Go 7400 (aka Quadro NVS 120M) in a new Dell Lattitude D820 also gave erroneous scores, but hand testing gave surface rendering score 420 and horrible mesh (110) and solid (<20) numbers. Tom Jeff wrote: > could you run the benchmark on your mac? (I have run it on my > workstation, but forgot the scores; will do again and compare) > > -jeff

1 0

NumPy in Chimera
by Tom Goddard 07 Feb '07

07 Feb '07

Hi Chimera Extension Writers, If you have written Chimera extensions that use Numeric Python you will need to convert them to instead use NumPy to work with future versions of Chimera. NumPy provides almost all the same routines as Numeric but some function and module names have changed. We are switching Chimera over from Numeric Python to NumPy and will have a Chimera snapshot out in one to three weeks that uses NumPy. We envision having a Chimera production release that uses NumPy in perhaps 3 months. The change was necessary because Numeric is no longer maintained. Numeric will no longer be included with Chimera because dynamic library name space collisions with NumPy occur on the Mac because of certain Chimera compilation options. The conversion of Chimera code was not much trouble, but we had to also convert third party packages that are included with Chimera (PyOpenGL, MMTK, Scientific (for netcdf file reading), MSMS molecular surfaces) and that took some work. Thanks for your patience. Tom

1 0

Re: [Chimera-users] question about mac laptops / workstations and chimera stability
by Tom Goddard 06 Feb '07

06 Feb '07

Hi Jeff, I just got my MacBook Pro with 2 Gb memory and 160 Gb hard-drive yesterday. It's been working well with Chimera volume display, surface and solid. Performance is roughly a factor of two better than my 4 year old Dell Precision M50 with Quadro 4 GoGL 500 graphics (a very heavy machine, ~10lb). I just took a look at a 2k by 2k by 76 tomogram and performance was as expected. 2 Gb memory is adequate for my uses. I went for the 160 Gb harddrive instead of the 120 Gb since volume data and video can take some space. The 200 Gb option was a 4200 rpm drive which I suspect would be slower than the smaller 5200 rpm models so I avoided it. If you'd like to try my laptop before you order you are welcome to come over to my house in SF. Tom Jeff wrote: > Tom, > > is the base 2 GB memory sufficient, and do you have any reservations > regarding transfer speed from the harddrive? > > I am going to place an order, wanted to check with you first about > those 2 things before going ahead. > > (I figure, if you are developing on it, this is not a bad idea.) > > -Jeff

1 0

Map Orientation for Symmetry Plug-in
by Albion Baucom 06 Feb '07

06 Feb '07

I am using the Symmetric Molecule Copies plug-in with Chimera. I am wondering if there is a preferred way to orient an EM volume so that it is centered around the z-axis at the origin. From my experience with this plug-in, symmetry related molecules are generated around z-axis at the origin of the viewing area. I am finding that my maps tend to be close, but not exactly oriented around the z-axis for a helical volume. Molecules fit into a portion of the map then tend to have slightly askew symmetry mates relative to the volume (presumably because the z-axis of the map is not exactly oriented along the origin of the symmetry z-axis). After playing around with various map tools (MAPMAN, Priism), I am still not satisfied with my orientation protocol and wonder if anyone can suggest a tried and true method for this. Thanks Albion

2 1

Re: [Chimera-users] volume/3586 problem zoning out volume
by Thomas Goddard 06 Feb '07

06 Feb '07

Hi Jeff, To check the memory limits set by a bash shell use shell command "ulimit -a". For a csh use "limit". Here's an example on my Mac (no memory limit). Tom $ ulimit -a core file size (blocks, -c) 0 data seg size (kbytes, -d) 6144 file size (blocks, -f) unlimited max locked memory (kbytes, -l) unlimited max memory size (kbytes, -m) unlimited open files (-n) 256 pipe size (512 bytes, -p) 1 stack size (kbytes, -s) 8192 cpu time (seconds, -t) unlimited max user processes (-u) 100 virtual memory (kbytes, -v) unlimited Jeff wrote: > I had forgotten about 4-byte floats versus 2-byte integers, > > you are right in that this is most likely a machine limitation, I was > surprised to get a fault and error about an allocation call. is there a > specific environment variable I should check to see if the shell is > limiting chimera's memory requests, I will change that if possible (?). > > -Jeff >

1 0

rotation by a set angle
by Eva Vanamee 05 Feb '07

05 Feb '07

Hi, I am trying to make a figure of my EM fit at different orientations. Is it possible to rotate the molecule/map by a certain angle so I know exactly the rotation between views? Thanks in advance, - Eva

3 2

Re: [Chimera-users] not able to 'transfer markers'
by Thomas Goddard 05 Feb '07

05 Feb '07

Hi Jeff, Menu entry "Actions / Transfer selected markers to current set" in the path tracer dialog moves the markers to the marker set chosen on the first line of the path tracer dialog. That should be the same as the one chosen in the marker set scrolling list (Features / Marker set list) unless that list has multiple sets selected. The error message "Markers in one set cannot link to markers in a different set" means you tried to transfer some markers to the current set but they were connected by at least one link to markers in another set. That is not supported. There is a command, keyboard shortcut "sc" that extends the current selection to all markers connected to currently selected ones. You'll need to turn on keyboard shortcuts with main window menu entry Tools / General Controls / Accelerators On. You can have them turned on by default every time you start Chimera using the Preferences dialog (menu Favorites / Preferences) category Tools and clicking the check button in the autostart column next to "Accelerators On". One possible confusion about hiding/showing markers is that if you have any markers selected then the Actions / Show & Hide menu entries act on those selected ones no matter what marker set they belong to. If there are no markers selected then those same menu entries act on all markers in the current marker set, that is the one chosen at the top of the path tracer dialog. The marker set scrolled list is used for closing multiple marker sets and for saving multiple marker sets. This may have been the problem when you could not close a marker set. It closed the one selected in the scrolled list which was different from the "current marker set" shown at the top of the path tracer dialog. When you change the current marker set menu it unfortunately does not clear the selected marker sets in the scrolled list. That is a bug. Let me know if you can reproduce the problem hiding/showing markers. Tom Jeff wrote: > hi Tom, > > using build 2318, if I select a series of markers, and then click > 'actions -> transfer selected markers to current set', it gives the > error 'Markers not transfered. Markers in one set cannot link to markers > in a different set.' when I did this, I assumed that the 'current set' > meant the one selected in the 'marker sets' dialog box in the volume > path tracer. > > is there a way for me to get around this? I had traced out a series of > actin filaments in one set, and for practical reasons want to be able to > show/hide each marker path that corresponds to a filament; the only way > I could think of doing this was to have each filament be a separate > marker file, unless there is a command that allows me to select a single > marker and then just 'display markers connected to selected marker' > (preferred, actually), as opposed to using the up-arrow, which selects > the entire set. > > after performing the 'transfer markers' request, I was no longer able to > show/unshow the marker sets. Also, when I closed a marker set, it was > still displayed. I shut down the session and recreated it from scratch. > I'll let you know if this show/unshow/close marker file issue occurs > again and under what circumstances; it would be hard for me to claim > that it was correlated to the 'transfer markers' command, I just know > that after that point in the session I could no longer show/unshow/close > the markers. > > thanks, > > -Jeff >

1 0

PDB file format for movie
by Mingfeng Yang 05 Feb '07

05 Feb '07

Dear Chimera users and developers, I'd like to use chimera to view the motions of a protein. If all the snapshots are to be saved in a single PDB file, what's the correct format for them? I tried the following, but chimera complained that a single structure is included only. Thanks, Mingfeng ------------------ MODEL 1 ATOM 1 N ALA 1 -4.564 -7.956 -6.531 1.00 0.00 N ATOM 2 CA ALA 1 -5.237 -8.746 -5.532 1.00 0.00 C ATOM 3 C ALA 1 -4.360 -9.172 -4.378 1.00 0.00 C ... ATOM 81 HN ALA 10 0.195 11.398 3.736 1.00 0.00 H END MODEL 2 ATOM 1 N ALA 1 -5.031 -8.234 -5.174 1.00 0.00 N .... END

3 3

Re: [Chimera-users] question about mac laptops / workstations and chimera stability
by Thomas Goddard 05 Feb '07

05 Feb '07

Hi Jeff, I have not seen any graphics driver bug reports with the MacBook Pro with X1600 graphics. I just ordered this machine to replace my four year old Dell with Quadro 4 graphics -- I am confident it will work well using volume data in Chimera. Here is our graphics driver bug page of problems observed with Chimera http://www.cgl.ucsf.edu/chimera/graphicsbugs.html It only has about 1/3 of the driver bugs and I will be adding more from our Chimera bug tracking system when I have time. Tom

1 0

volume path tracer / render surface with marker set
by Jeff 02 Feb '07

02 Feb '07

hello, is there a way to render a surface of arbitrary thickness using a marker set? as in, fit some sort of spline to a marker set, with the constraint being that the algorithm assumes they all lie on the same surface? this would be useful for modeling membranes, perhaps it is already implemented and I am not seeing it in the dialogs. thanks, -J

2 1