NVIDIA Launches Blackwell Platform to Empower Next-Generation Computing

According to PANews, NVIDIA announced the launch of the NVIDIA Blackwell platform to empower the new era of computing. The platform enables institutions worldwide to build and run real-time generative AI on large language models (LLM) with trillions of parameters, reducing costs and energy consumption by up to 25 times compared to the previous generation. The Blackwell GPU architecture features six revolutionary acceleration technologies that will drive breakthroughs in data processing, engineering simulation, electronic design automation, computer-aided drug design, quantum computing, and generative AI. The NVIDIA GB200 Grace Blackwell superchip connects two NVIDIA B200 Tensor Core GPUs with the NVIDIA Grace CPU through a 900GB/s ultra-low-power interconnect. Compared to the NVIDIA H100 Tensor Core GPU, the GB200 superchip achieves up to 30 times performance improvement in LLM inference workloads, with cost and energy consumption reduced by up to 25 times. In addition, NVIDIA also introduced a general-purpose large model called GR00T for the robotics field during the GTC conference in San Jose, USA. NVIDIA also launched a new computer called Thor, specifically optimized for robotics in terms of performance, power consumption, and size. NVIDIA aims to give robots smarter brains through this initiative, enabling them to naturally imitate human behavior by observing it, greatly improving coordination and flexibility in their actions.

查看原始新闻