Techware Labs Header
RSS
Home | Reviews | Articles | Downloads | Guides | Staff | Advertising | Links
Mainboards | Networking | Video | Cases | Storage | Other

CUDA, What Is It and What Does It Mean To You?


Author:  Chris Swertfeger
Date:  2008.09.25
Topic:  Editorials
Provider:  Palit
Manufacturer:  Palit






CUDA, What Is It and What Does It Mean To You?

Jetart

Jim, I'm a Doctor Not a Programmer:

Obligatory Star Trek reference aside, for those that aren't 1337 h4x0r programmers, Nvidia's CUDA does have things to offer the common user, including a way to help cure cancer.

Jetart

If you would lke to join the techwarelabs folding at home team,

enter team number "147447" in the folding at home program.

Folding@Home is a distributed computing effort created by Stanford University to help understand protein folding, and in turn understand how certain diseases form. The program they offer takes advantage of idle CPU time (most computers don't fully utilize their CPU at all times). All of the extra processing time not being used is then harnessed to help calculate protein folding. Due to the nature of distributed computing projects, they are ideal for massive amounts of processing cores, something that the GPU has. On August 1, 2008 Stanford (with Nvidia's help) released a CUDA enabled Folding@Home client for use with their new GPU's. Since its release, Nvidia GPU's have contributed more processing power than any other group of processing units (contributing over 1 PetaFLOP of processing power). This has great potential for the medical field and for many other distributed computing efforts that want to harness this kind of power.



OS Type Current TFLOPS Active CPUs Total CPUs
Windows 208 218542 2152602
Mac OS X/PowerPC 7 8345 119278
Mac OS X/Intel 21 6672 60368
Linux 60 35357 327844
ATI GPU 414 3766 7462
NVIDIA GPU 1403 12758 21540
PLAYSTATION 3 1236 43816 590161
Total 3349 329256 3279255

As of Aug 28, 2008 06:04:43

What about the gamers?:

Some may remember the company Ageia and their PhysX card that claimed to offer realistic physics in games without impacting performance. In February of this year, Nvidia bought out Ageia and with them, their PhysX engine. Now dubbed Nvidia PhysX, it offers all the same features of the old PhysX with one huge advantage. It can be run on any CUDA enabled gaming graphics card, no separate PhysX card required. That means that the only thing you need to enjoy realistic physics in supported games is a supported graphics card and the latest drivers.

Nvidia offers a mod pack for the game Unreal Tournament 3 on their site that acts as a tech demo for PhysX. Some of the features of the levels in the mod pack include:

REALTIME PRICING

 

 



« How Can I have CUDA?
But what if I own an ATI card... »