Cerebras Systems , a pioneer in generative AI acceleration, and G42 , a United Arab Emirates-based technology group, unveiled Condor Galaxy, a network of nine interconnected supercomputers that offers a new approach to AI computing. To do this, it promises to significantly reduce the training time of AI models. Condor Galaxy 1 (CG-1), the first AI supercomputer on this network, has 4 exaFLOPs and 54 million cores. Cerebras and G42 plan to deploy two more such supercomputers, CG-2 and CG-3, in the US in early 2024. With a planned capacity of 36 exaFLOPs in total, this unprecedented supercomputing network will revolutionize the advancement of AI globally.
“Partnering with Cerebras to rapidly deliver the world’s fastest AI training supercomputer and laying the groundwork for interconnecting a constellation of these supercomputers around the world has been extremely exciting. This alliance brings together Cerebras’ extraordinary computing capabilities, along with G42’s cross-industry AI expertise. G42 and Cerebras share the vision that Condor Galaxy will be used to address society’s most pressing challenges in healthcare. , energy, the fight against climate change, etc.”, says Talal Alkaissi, CEO of G42 Cloud, a subsidiary of G42.
Also Read: Mastek to Strengthen Data Cloud and Generative AI Capabilities with Acquisition of BizAnalytica
Located in Santa Clara, California, CG-1 links 64 Cerebras CS-2 systems into a single, easy-to-use AI supercomputer, with an AI training capacity of 4 exaFLOPs. Cerebras and G42 offer CG-1 as a cloud service, allowing customers to enjoy the performance of an AI supercomputer without having to manage or distribute models on physical systems.
The CG-1 marks the first time that Cerebras has partnered not only to build a dedicated AI supercomputer, but also to manage and operate it. The CG-1 is designed to enable G42 and its cloud clients to train large innovative models quickly and easily, thereby accelerating innovation. The Cerebras and G42 strategic alliance has already advanced cutting-edge AI models in bilingual Arabic chat, healthcare and climate studies.
“With 4 exaFLOPs of AI computation at FP 16, the CG-1 dramatically reduces AI training times and eliminates the pain of distributed computation,” says Andrew Feldman, CEO of Cerebras Systems. “Many cloud computing companies have announced massive GPU clusters that cost billions of dollars to build but are extremely difficult to use. Distributing a single model across thousands of tiny GPUs requires months of time from dozens of people with unusual knowledge. CG-1 removes this obstacle. Setting up a generative AI model takes minutes, not months, and can be done by one person alone. The CG-1 is the first of three 4-exaFLOP AI supercomputers to be used in the US along with G42,
SOURCE: Businesswire