Nvidia Reveals How Artificial Intelligence Can Improve Computer Chip Design

Nvidia chief scientist, Bill Dally, said artificial intelligence could improve chip design. (photo: doc. nvidia)

JAKARTA - Leading chip designer company, Nvidia Corp, on Monday, March 27 unveiled new research explaining how artificial intelligence (AI) can be used to improve chip designs. The chip design process involves deciding where to place tens of billions of tiny switches called transistors on a single piece of silicon to make a working chip.

The exact placement of transistors has a major impact on the chip's cost, speed, and power consumption. Chip design engineers use complex design software from companies such as Synopsys Inc and Cadence Design Systems Inc to help them optimize transistor placement.

On Monday, Nvidia released a paper showing that they could use a combination of artificial intelligence techniques to find better ways to place large groups of transistors.

The paper aims to improve on Alphabet Inc's 2021 paper by Google, which has since become the subject of controversy. Nvidia's research took an existing effort developed by the University of Texas researchers using so-called reinforcement learning and added a second layer of artificial intelligence on top of it to get even better results.

Nvidia chief scientist Bill Dally said the work was important because improvements in chip manufacturing were slowing with the cost per transistor in the new generation of chip manufacturing technologies now being higher than in previous generations.

That contradicts the famous prediction by Intel Corp. founder Gordon Moore that chips will always get cheaper and faster. "You're not getting any economics out of that scalability anymore," Dally said, as quoted by Reuters. "To move forward and deliver more value to customers, we can't get it from cheaper transistors. We have to get it by getting smarter in design."