AI Research Super Cluster
Image Source : ai.facebook.com/blog
Meta has launched its new AI Research Super Cluster (RSC) supercomputer, designed for accelerating AI research. The supercomputer, which has been in the works for two years, could finally help Meta develop much more powerful AI software — potentially useful for complex tasks like detecting hate in posts.
Supercomputer includes 6,080 GPUs split across 760 nodes, potentially making it one of the most powerful in the world. The AI Research Super Cluster is an AI supercomputer that will be the world’s largest and fastest when it is fully built.
Facebook has long been betting that artificial intelligence can help it with the daunting task of moderating posts from billions of users. Artificial intelligence requires that a computer perform large numbers of calculations simultaneously with low precision. This task is best suited for GPUs, with thousands of cores running simultaneously. Conventional supercomputers can be optimized for high-precision operations. AI computers work at lower accuracy levels, which allows for significant increases in computational speed without affecting final results. A supercomputer is a high-performance infrastructure.
It is necessary to process the large amount of data Meta has. According to reports, the Meta supercomputer is being used for research purposes. It is unlikely that products built on it will be available in the future. The company’s goal is to develop AI models that work and human brains and provide context understanding. They will analyze text, images and videos in hundreds of languages and create augmented reality tools.
This technology will allow Meta to identify malicious content faster and more accurately. It also allows its researchers to develop AI models that think like humans and provide rich multi-dimensional experiences in the metaverse.