[Beowulf] [openworm-discuss] Accelerating compartmental modeling on a graphical processing unit

Mark Hahn hahn at mcmaster.ca
Wed Mar 20 16:04:57 PDT 2013


On Wed, 20 Mar 2013, Eugen Leitl wrote:
> ----- Forwarded message from Mike Vella <vellamike at gmail.com> -----
...
> http://www.frontiersin.org/Neuroinformatics/10.3389/fninf.2013.00004/abstract

thank you for forwarding that.  it's a very nice paper, with useful
and practical comments on Cuda coding.

it's a bit unfortunate that people doing Cuda coding can never seem 
to resist the urge to compare their work versus absurd strawmen.
ie, usually code that hasn't been as carefully tuned for CPUs, 
and/or almost always against one core of an old/slow CPU.

Tesla C2075 versus a circa 2009 PhenomII x4 920!


More information about the Beowulf mailing list