IBM Storage Scale System 6000 Certified Nvidia Cloud Partner

By Mike Kieran, HPC/AI storage marketing, IBM Corp.

IBM Storage Scale System 6000 is now a certified Nvidia Cloud Partner (NCP) for HGX  H100/H200/B200 systems.

As a certified high performance storage partner for NCP, IBM Storage Scale System 6000 has demonstrated that it can deliver scalable high-performance IO to the demanding AI training and inferencing workloads deployed on HGX GPUs in the cloud. 

Storage Scale System 6000 is a hardware platform that’s designed to be the simplest and fastest way for organizations to build a global data platform around their file and object data. It leverages the power of Storage Scale software combined with NVMe flash and hybrid flash/disk technology to deliver high-performance storage for AI, data analytics, and file and object use cases.

The Nvidia HGX AI supercomputing platform combines Nvidia GPUs, NVLink switches, Nvidia networking, and fully optimized AI and high-performance computing software stacks, designed to work together to maximize application performance. It’s for customers running AI, complex simulations, and massive datasets that require multiple GPUs with extremely fast interconnections.

Storage Scale System 6000 is available in all-flash and hybrid configurations providing:

• Up to 310GB/S throughput with low latency;
• Up to 13 millions IOs using NVMeoF;
• Up to 1.8PBe (effective capacity) in a standard 4U rack space.

Resource:
IBM Scale System 6000 certifications for Nvidia GPU systems