NVIDIA CUDA to deliver the whoop ass

NVIDIA VP of Content Relations Roy Taylor says the world will soon discover just how pathetic conventional processors really are. Taylor says NVIDIA's CUDA programming platform will allow software developers to unlock the full potential of GPUs. NVIDIA reckons graphics chips can perform certain tasks 20x or even 100x faster than today's multi-core processors.

1 TFLOP GPU
TechRadar points out that the University of Reading just installed a new room-sized high density 20 TeraFLOPS computing and compares this to NVIDIA's new GPU that will be released this summer. This chip is rumoured to have a raw computing power of 1 TFLOPS and Tech Radar believes a four-way GPU node could deliver the same computing power as the new Reading University supercomputer:

On paper, it's extremely plausible. In terms of raw parallel compute power, 3D chips put CPUs to shame. A good recent example is the new room-sized, high density computing cluster installed by Reading University.

Designed to tackle the impossibly complex task of climate modelling, it weighs in at no less than 20 TeraFlops. That sounds impressive until you realise that just a single example of Nvidia's next big GPU, due this summer, could deliver as much 1TFlop. So, a few four-way Nvidia GPU nodes will soon offer the same raw compute power as a supercomputer built using scores of CPU-based racks.

Elemental HD - superfast video encoding
One of the first major applications of NVIDIA's CUDA technology for consumers will be video encoding. The GPU maker is developing a video encoding application known as Elemental HD and says a GeForce 8800 graphics card will be able to downsize a typical HD movie for an iPod in just over 20 minutes. That may sound long but a decent dual-core Intel processors may need up to eight hours of more to finish this job!

Downsizing a typical HD movie for an iPod using a conventional PC processor can take up to eight hours or more, even with a decent dual-core Intel chip. Nvidia says the same job can be done in just over 20 minutes on an 8800 series Nvidia graphics board.

"When you look at the question of whether you should transcode video on a GPU or CPU, when you consider it in performance-per-buck terms, it's currently obscenely the wrong way round," Taylor says. And the solution is simple enough. Don't spend any more money overall. Just spend a little less money on your Intel CPU and a little more on your Nvidia GPU.

Besides video processing NVIDIA is also working on a CUDA version of AGEIA's PhysX processor. A software download that will add PhysX support to GeForce 8 and 9 series graphics cards is expected this Summer.