Meta’s Next-Gen ‘AI Research Super Cluster’ will Revolutionize Artificial Intelligence
Currently, AI can carry out tasks such as translating languages, helping identify potentially harmful content, etc. However, developing the next generation of AI will require much more powerful supercomputers that are capable of quintillions of operations within one second. As per a source, this supercomputer is being developed by connecting 6,080 state-of-the-art A100 GPUs of the Nvidia company. The company is further planning to add another 10,000 GPUs in the upcoming months, which would make it the world’s most powerful supercomputer in the field of artificial intelligence and machine learning. Even in its contemporary state, this supercomputer can transmit data at speeds of up to 200 GBs per second. If we talk about memory, the supercomputer has the enormous capacity to store as much as 175,000 terabytes of data. Its cache is 46000 terabytes of data, while under the Network File System (NFS), it can store as much as 10,000 terabytes of data. Some trials were carried out to check the performance of the AIRSC. Using its extraordinary power of computing on big data, AIRSC recognizes objects embedded in the photos and scenes, identifies particular words in conversations, interprets misinformation and hatred on social media, etc. Furthermore, the supercomputer would also be able to carry out tasks such as rapid training of algorithms that identify provocative content on online platforms. Check out? Apple AirPods Pro 2 to be Released in Third Quarter of the Year
