B-100 and X-100 Nvidia AI Chips and Zettaflops AI Compute Clusters
by Brian Wang from NextBigFuture.com on (#6HK03)
The NVIDIA H100 GPU has 30 teraflops of peak standard IEEE FP64 performance, 60 teraflops of peak FP64 tensor core performance, and 60 teraflops of peak FP32 performance. It also has a second-generation secure multi-instance GPU technology that makes it possible to partition the chip into seven smaller, fully isolated instances to handle multiple workloads ...