Inspur released a high-performance computing system (HPC) using Intel Xeon Platinum 9200 series processors with 112 cores and up to 9.3 TFLOPS of Floating Point 64 performance per node.
It is designed for massive parallel HPC and AI applications, image and video processing, and virtrual apps, the company said on Friday.
Inspur’s cluster management platform called ClusterEngine will be combined with Teye, an HPC application performance monitoring and analysis system. There are 24 memory channels per node and on-board liquid cooling.
The 9200 is the fastest processor in the second-generation Xeon Scalable processor family. Tests on a single node Intel Xeon Scalable 9242 show that it can deliver memory bandwidth at 300 GB/s when running a weather application. That test used 24 32GB of 2933MHz memory.
Inspur is one of the world’s top three server manufacturers.
The Inspur platform was shown at Supercomputing 2019 in Denver. Also at the show, Intel unveiled an exascale Ponte Vecchio GPU as part of the Aurora project at Argonne National Lab.