After Virtual SC20: Inspur Update Focuses on Staying Ahead in HPC and AI

Print Friendly, PDF & Email

In this interview, Inspur’s Global AI and HPC Director Vangel Bojaxhi sat down with us after virtual SC20 to give us an HPC- and AI-oriented update on the company. Noting that Inspur is a top three server manufacturer in the world, Bojaxhi added that the company is no. 1 in GPU servers.

At SC20, Inspur showcased two servers tailored for AI workloads, the NF5488A5 and NF5488M5-D, which provide support for the latest Nvidia A100 80G GPU and are now available globally. Earlier this year, Bojaxhi pointed out, the NF5488A5 server set a new AI performance record in the Resnet50 training task, topping the MLPerf AI benchmark list for single server performance.

“We plan to … integrate the latest technologies expected from our main partners soon to deliver to our customers a richer AI ecosystem and AI product portfolio of super-hybrid server architecture design.”

At the conference, Inspur also highlighted AIStation, a deep learning management system built for AI training clusters that supports several types of deep learning frameworks. AIStation unifies the management, scheduling and monitoring of CPU and GPU resources during training runs.

Looking ahead to 2021 and 2022, Inspur will build out its cloud-based HPC-as-a-service portfolio and, in AI, emphasize server inferencing capabilities, demand for which is expected to exceed that for AI training by 2022, Bojaxhi said.

“We plan to align with this next year,” he said, “and integrate the latest technologies expected from our main partners soon to deliver to our customers a richer AI ecosystem and AI product portfolio of super-hybrid server architecture design.”