CES: Nvidia licenses ARM tech for supercomputer chip

CES: Nvidia licenses ARM tech for supercomputer chip

Summary: Graphical chipmaker Nvidia will license processor architectures from chip architecture designer ARM for a new chip code-named Project Denver, which is targeted at supercomputers as well as desktops

TOPICS: CES, Emerging Tech

Nvidia will license chip architecture from ARM for a new chip, code-named Project Denver, that is targeted at high-performance computing, servers and desktop computers.

The chip will combine a CPU, based on an as-yet-unnamed ARM architecture, and an Nvidia GPU, Nvidia announced at the Consumer Electronics Show on Wednesday.

Consumer Electronics Show

CES 2011 coverage

Catch up with the latest technology unveiled at CES 2011 in Las Vegas.

Read more

"Denver will usher in a new era for computing by extending the performance range of the ARM instruction-set architecture, enabling the ARM architecture to cover a larger portion of the computing space," Nvidia's chief scientist Bill Dally wrote in a blog post. "Project Denver extends the range of ARM systems upward to PCs, datacentre servers and supercomputers."

Denver should be more energy efficient than x86 processors, Dally said, because of its wider instruction set.

Nvidia will also license ARM's Cortex-A15 processor architecture for its future Tegra mobile processors. The company has been licensing the Cortex-A9 architecture for its current Tegra 2 chips.

Denver is not the first time that an ARM-based chip has been considered for use in servers. In May, ARM announced it was running one of its websites on a cluster of ARM chips, with other trials under way. In November chipmaker Marvell announced the Armada XP, a quad-core ARM-based processor intended for servers.

New opportunities
ARM will not comment on speculation about what type of architecture will be licensed to Nvidia for Project Denver, ARM marketing chief Ian Drew told ZDNet UK on Thursday.

"Clearly it is a special thing in that [Nvidia] are one of the first people to do it, so we're really pleased they've chosen our architecture to go into this space," Drew said.

"It's not an easy thing to do, either. If you think of all the server architectures in the world, if you think of the complexity of the server and the software requirements and the hardware requirements, it's not just taking a chip out of a phone and putting it into a server — there's a whole lot more complexity than that," he added.

Drew said that the expansion of ARM from its low-powered device roots created new opportunities. "I look at the way phones have developed over the [last 10 years], from purely voice to WAP to 2.5G to 3G to getting my emails on them, to getting my apps on there, to tablets coming out," he said. "I think our business model has allowed the hardware guys and the SoC [system-on-a-chip] guys to differentiate their products and some of them work and some of them don't, but it allows a gamut of products to be brought out that were not there before."

Get the latest technology news and analysis, blogs and reviews delivered directly to your inbox with ZDNet UK's newsletters.

Topics: CES, Emerging Tech

Jack Clark

About Jack Clark

Currently a reporter for ZDNet UK, I previously worked as a technology researcher and reporter for a London-based news agency.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


1 comment
Log in or register to join the discussion
  • The NVIDIA GPUs that are used in video cards and Tesla compute card have hundreds of processor cores which can run supercomputer programs. It would be interesting if NVIDIA could combine ARM architecture (ARM doesn't make the chips its just licences designs) with its GPU cores and have hundreds of combined instruction set cores on a single processor chip. A Supercomputer on a chip. Also a combined Arm processor core and Nvidia GPU on a chip could make a better tablet or phone. Sounds exciting.