Hi,

We have been developing tools that can leverage GPU technology. Have a look at this:

http://cci.lbl.gov/cctbx_sources/cudatbx/

Basically, the infracstructure is available for building the CCTBX with a cuda compiler and calling those functions from within python using the basic boost python wrappers.

There is not much there in the cudatbx yet, but in some other projects we do have direct summation code that runs very nicely (not crystallographic applications). As Nat mentioned however, it is unclear what the total runtime impact would be though.

Cheers
Peter




On 10 January 2012 11:56, Nathaniel Echols <nechols@lbl.gov> wrote:
On Tue, Jan 10, 2012 at 11:21 AM, Schubert, Carsten [JRDUS]
<CSCHUBER@its.jnj.com> wrote:
> are there any development plans for phenix in the works to take advantage of
> NVIDIA’s CUDA platform?

I was wondering when someone would ask about this.  Short answer is
that there are cases where CUDA (and/or OpenCL) might be useful*, but
it is unlikely to have a significant impact on the runtime of most
existing programs in Phenix.  Various other algorithmic improvements
are more likely to have a substantial impact on speed than optimizing
our code for fancy hardware, especially for phenix.refine.

-Nat

(* and some work is actually being done on this in CCTBX, but not as
part of Phenix per se.)
_______________________________________________
phenixbb mailing list
phenixbb@phenix-online.org
http://phenix-online.org/mailman/listinfo/phenixbb



--
-----------------------------------------------------------------
P.H. Zwart
Research Scientist
Berkeley Center for Structural Biology
Lawrence Berkeley National Laboratories
1 Cyclotron Road, Berkeley, CA-94703, USA
Cell: 510 289 9246
BCSB:      http://bcsb.als.lbl.gov
PHENIX:   http://www.phenix-online.org
SASTBX:  http://sastbx.als.lbl.gov
-----------------------------------------------------------------