A computing chip that uses a large amount of electricity by artificial intelligence. On Wednesday, the world's largest manufacturer of these chips unveiled a new strategy to make them more energy efficient. Design them using AI-powered software.
At a meeting in Silicon Valley, Taiwan Semiconductor Manufacturing Company (TSMC)the contract manufacturer of Nvidia's chips introduced various ways they wanted to increase the energy efficiency of AI computing chips about 10 times.
For example, Nvidia's current flagship AI server can consume up to 1,200 watts during demanding tasks.
GAINS TSMC hopes to achieve from a new generation of chip designs, packaged with different technologies, multiple “chips” (smaller computing chips) to create a single computing package.
However, to utilize these technologies, companies designing chips are increasingly relying on AI-powered software from providers. Cadence Design System and SynopsysBoth launched new products developed in close collaboration with TSMC on Wednesday.
In some of the complex tasks of designing chips, the tools of TSMC's software partners have found better solutions than TSMC's own human engineers.
