In the middle of last year, NVIDIA announced Grace, its first data center processor. At the time, the company only shared a few details about the chip, saying, for example, that it would use its NVLink technology to provide data transfer speeds of up to 900 GB/s between components. Now, according to Engadget, the company’s CEO, Jensen Huang, has introduced the Superchip CPU Grace, the first discrete CPU that NVIDIA plans to release as part of its Grace lineup.
Based on ARM’s v9 architecture, the Grace CPU Superchip is actually two Grace CPUs connected via the company’s NVLink interconnect technology. Includes the staggering number of 144 ARM cores in a single socket and consumes approx. 500 watts of power. Ultra-fast on-chip LPDDR5x memory enables bandwidth speeds of up to 1 terabyte per second.
Although they are very different chips, a useful way to conceptualize NVIDIA’s new device is to think of the recently announced Apple M1 Ultra. In simpler terms, The M1 Ultra is made up of two M1 Max chips connected through Apple’s UltraFusion technology.
When NVIDIA begins shipping the Grace CPU superchip to customers like the Department of Energy in the first half of 2023, it will offer them the option to configure it as a stand-alone CPU system or as part of a server. with up to eight Hopper-based GPUs. The company claims its new chip is twice as fast as traditional servers.. NVIDIA estimates it will achieve a score of approximately 740 points on the SPECrate®2017_int_base benchmark tests, putting it in the top echelon of data center processors.