Google has announced its new A3 cloud supercomputer, which is now available in private preview. The A3 uses the Nvidia H100 GPU, Google’s custom-designed 200 Gbps VPUs, and Google’s Jupiter data center, which can scale to tens of thousands of interconnected GPUs. The A3 provides up to 26 exaFlops of AI performance and a 30x inference performance boost over the A2. It also features 4th Gen Intel Xeon Scalable processors and 2TB of host memory via 4800 MHz DDR5 DIMMs.
Previous ArticleIntroducing Mediapipe Solutions For On-device Machine Learning
Next Article 3 Questions: Jacob Andreas On Large Language Models