Nvidia Unveils Hopper H100 GPU and Grace Megachip for HPC
Nvidia introduces its new Hopper H100 GPU and Grace CPU superchips alongside updated DGX systems to accelerate high-performance computing and AI workloads.
Nvidia officially launches its highly anticipated Hopper H100 GPU, marking a significant leap in processing power for artificial intelligence and high-performance computing. The new architecture features a dedicated Transformer Engine designed specifically to accelerate large language models and deep learning tasks.
Alongside the new GPU, Nvidia reveals the Grace megachip, which combines two Grace CPUs into a single superchip to handle complex data center workloads. The company also introduces updated DGX systems that integrate these new processors to provide massive performance improvements over previous generations.
These new hardware offerings represent Nvidia's aggressive push to dominate the exascale computing era. By delivering unprecedented speed and efficiency for AI training and scientific simulations, Nvidia equips researchers and enterprises with the tools needed to tackle the world's most complex computational challenges.