On Mon, Jan 24, 2005 at 02:33:35PM -0500, Dan Goe wrote: > What about PVM(Parallel Virtual Machine) or MVS ?
Shared-memory is an expensive (exponentially so, if you want to maintain coherency of N spatially separated instances of same data in write-intensive context) mirage. But sure, you can combine GPUs with MPI, assuming you can make your problem spatial-tesselation, exchange-only-surfaces-of-adjacent-volume-regions. Most physical simulations can be formulated in this way, and the CNS is just an anisotropic nonlinear excitable 3D medium as far as physics is concerned. > > Dan Goe > > > ----- Original Message ----- > From: "Eugen Leitl" <[EMAIL PROTECTED]> > To: <[email protected]> > Sent: Monday, January 24, 2005 2:59 AM > Subject: [agi] RE: non-traditional processors or GPUs (fwd from > [EMAIL PROTECTED]) > > > ------- > To unsubscribe, change your address, or temporarily deactivate your > subscription, please go to > http://v2.listbox.com/member/[EMAIL PROTECTED] -- Eugen* Leitl <a href="http://leitl.org";>leitl</a> ______________________________________________________________ ICBM: 48.07078, 11.61144 http://www.leitl.org 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE http://moleculardevices.org http://nanomachines.net ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
pgpiGhotMFmUe.pgp
Description: PGP signature
