Hi Pascal,

  I'd suggest having 8 Gbytes of memory, although 16 Gb would be better.  With big maps (say 500 x 500 x 500) grid points and fitting a few hundred thousand atoms you can use over 4 Gbytes of memory and performance can get poor if you have too little memory for all the apps running on your machine. Chimera is single-threaded so multiple cores won't help much, except to avoid resource contention with other running applications.  On-board intel graphics is quite reasonable for medium size maps and tens of thousands of atoms, but for your larger systems an Nvidia or AMD gaming type mid-range graphics cards ($100) will be much (2 -  5 times) faster to rotate and move the models.  This web page shows how well various graphics cards perform with Chimera

http://plato.cgl.ucsf.edu/trac/chimera/wiki/benchmarks

If you wanted some really fast graphics, a Geforce GTX 970 for $300 is a good deal.

http://www.videocardbenchmark.net/high_end_gpus.html

  Tom


On Mar 15, 2016, at 2:54 AM, Pascal Albanese wrote:

Dear staff,
we need some advice about the computational power required by the software to perform several operations/processing with big volumes obtained by cryoEM at high resolution. For example the fitting of several pdb crystal structure (n=20-30) into a single volume at high resolution of a protein complex (MW > 1.2 MDa).
What kind of graphics power do we'll need to run Chimera properly with this kind of structures? and CPU/RAM performance?
If you have any suggestion.

Thanks in advance for your help,

all the best,

Pascal

--
Pascal Albanese, 
PhD student, Università di Padova
Via Ugo Bassi 58 B, 35121 Padova, Italy

Research assistant, Politecnico di Torino
Dipartimento DISAT, BioSolar Lab
Viale Teresa Michel 5, 15121 Alessandria, Italy

Tel: 0131 229301 Fax: 0131 229344

_______________________________________________
Chimera-users mailing list: Chimera-users@cgl.ucsf.edu
Manage subscription: http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users