Does NVIDIA Selene Form A Wider Moat Than CUDA?

The annual International Supercomputer Conference (ISC), held virtually this year, kicked off today. Not surprisingly, NVIDIA has already made a few announcements of note. Especially of interest to me was the announcement of Selene, NVIDIA’s in-house 1+ Exaflop AI supercomputer, which ranks as the fastest industrial system in the USA and #7 overall in the Top 500. . NVIDIA also announced a new PCIe version of the A100 accelerator, six A100-based supercomputer wins and a new Mellanox UFM Cyber AI platform to predict and detect security threats and predict network failures. Still, Selene was the star of the show.

Selene: A deep competitive moatost people think of CUDA when someone mentions NVIDIA’s competitive defenses. Certainly, the high-performance software is a significant advantage for NVIDIA, even 13 years after its introduction. CUDA enables HPC and AI applications to run efficiently on NVIDIA GPUs, and is embraced by programmers around the world. It supports thousands of applications on millions of GPUs. However, Selene may form an even more formidable defensive moat than the venerable CUDA libraries and tools.

Let’s look at Selene. It is comprised of 280 NVIDIA DGX A100 servers, each with 8 Ampere GPUs, interconnected by over 490 200Gb Mellanox Switches. Supercomputers typically require up to a year to be installed, but NVIDIA engineers assembled and tested the platform in under one month—a testament to the DGX platform’s plug-and-play ease of installation.

Back in 2017, NVIDIA announced the V100, along with the company’s Saturn V in-house supercomputer. A top-30 supercomputer built to enable research and development of NVIDIA software and hardware, the platform has been used to increase the performance of many AI and HPC workloads at scale….

Read More