Heart of the matter

The graphics processing unit or GPU is proving to be a challenge to the CPU in what has been called the ‘battle for the soul of the PC’. Recently, Nvidia’s CEO and co-founder Jen-Hsun Huang (Jensen) was in Dubai, and WINDOWS Gareth van Zyl caught up with him to discuss the future of computing, Nvidia’s focus in the region and why visual computing is becoming ever more important.

Tags: CPUGPUNVIDIA CorporationUnited Arab Emirates
  • E-Mail
Heart of the matter
More pics ›
By  Gareth Van Zyl Published  December 7, 2009

For Nvidia it all started out with the GeForce graphics cards for gamers. It's branched out into the likes of Tegra (which is incorporated into the Zune HD media players), Tesla (which is geared towards technical industries such as Oil and Gas), and CUDA, the architecture and computing engine that can now be found on many GPUs and which has helped to drive graphics development to new heights.

Of late, Nvidia's role in elevating the prominence of the GPU has never been more apparent than with the release of its ION platform, which works alongside low power central processing units, such as Intel's Atom processors, to allow for small form factor PCs to have better performance.

So, when Nvdia's co-founder and CEO, Jen-Hsun Huang (Jensen) recently visited the Middle East and was hosted at the American University of Dubai (AUD), it was a perfect opportunity to gauge his thoughts on the future of computing, on what Nvidia's interests in the Middle East are and why the GPU and visual computing are playing an increasingly important role in our digital lives.

On the future of computing, Huang starts off by saying,"In 1993, Nvidia's chip was one million transistors and a 46 microprocessor was four million transistors, so five million transistors would effectively represent the vast majority of a high-end CPU or high-end system. Well, today you have 3 billion transistors on Nvidia's new Firma chip."

The Firma chip is, of course, going to be incorporated into what will be Nvidia's ‘general purpose' GPUs. That is, turning GPUs into general purpose parallel computing processors that work alongside and assist CPUs to perform better.

"It's about number crunching," says Huang, "it's about the kind of problem that you have and whether it's possible to do it parallel or you have to do it in serial. If you can do something in parallel you would. So, we call this new world not central processing but co-processing. Instead of just using a sequential processor, there's going to be a sequential processor and a parallel co-processor and instead of just running everything on the CPU we're now going to have APIs like DirectCompute, Open CL and CUDA that allows people to write programs that takes advantage of the CPU and the GPU."

Add a Comment

Your display name This field is mandatory

Your e-mail address This field is mandatory (Your e-mail address won't be published)

Security code