Modular System Architecture for HPC and AI

HPC is a small market for technology providers, making it too expensive to develop products exclusively for this field. Instead, the high-end computer industry has traditionally developed products for the server market, from which they get their return of investment. We have used these off-the-shelf products to build HPC systems via cluster computing and parallel programming. Now, technology providers have moved their focus towards the cloud and, most recently, towards the exploding AI-market dominated by hyperscalers. The latter do dictate the design of (and even develop themselves) upcoming CPUs and accelerators, based on the requirements of their AI applications. To benefit from these new products, we in HPC must ask ourselves how the specific requirements of our traditional HPC applications differ from those of the most popular AI-models, and how we can make them compatible with each other or rather, how HPC systems should look like to serve the needs of both HPC and AI application domains. In this talk, we will present some of our observations on the similarities and differences between the requirements of HPC and AI applications, on the available processing technologies, and on the Modular Supercomputing Architecture that aims at serving both kinds of users.