Semiconductor firm Nvidia on Thursday announced a new chip that can be digitally split up to run several different programs on one physical chip, a first for the company that matches a key capability on many of Intel Corp’s chips.
Santa Clara, California-based Nvidia said its A100 chip help the owners of data centers get every bit of computing power possible out of the physical chips they purchase by ensuring the chip never sits idle.
When software developers turn to a cloud computing provider such as Amazon.com or Microsoft for computing power, they do not rent a full physical server inside a data center. Instead they rent a software-based slice of a physical server called a “virtual machine.”
Customers who want to test the theory will pay a steep price of $200,000 for Nvidia’s DGX server built around the A100 chips.
Nvidia CEO Jensen Huang said the computing power in the DGX A100 was equal to that of 75 traditional servers that would cost $5,000 each.
“Because it’s fungible, you don’t have to buy all these different types of servers. Utilization will be higher,” he said. “You’ve got 75 times the performance of a $5,000 server, and you don’t have to buy all the cables.”
Nvidia shifts strategy
Nvidia, whose semiconductors power data centers, autonomous cars and robots, said on Thursday it plans to enter the market for technology that helps cars with automated lane-keeping, cruise control and other driver-assistance features, Reuters reported on Thursday.
The move represents a change in direction for Nvidia. Until now, Nvidia has supplied key technology aimed at making autonomous vehicles that require much more sophisticated computers.
Danny Shapiro, senior director of automotive at Nvidia, said the shift in strategy is aimed at meeting the existing needs of automakers that struggle with maintaining two systems – one for the driver assistance available today, and one for more advanced self-driving technology for the future.
The new Nvidia system means automakers will be able to use one system for both, saving engineering efforts and using some of the self-driving technology to improve the driver assistance functions, Shapiro said.
Nvidia’s new self-driving technology uses the “Orin” processing chip the company launched in December. Shapiro said he expected vehicles using the system could start production in early 2023.
The Nvidia chips will be part of a larger system that includes cameras and will likely be built by traditional automotive suppliers such as Continental, ZF Friedrichshafen AG or Robert Bosch.