Forecast Update: GPU and Accelerator Growth in HPC

Historically, GPUs and accelerators have been growing roughly 26% per year (a five-year CAGR). Over the next five years, installations are expected to increase by at least 17% a year, greatly outpacing the revenue growth for HPC servers over the same forecast period. As AI workloads become an increasingly important part of overall HPC workload, both enabling new innovative techniques to solve new problems and to supplement traditional modelling and simulation workloads, GPU usage will likely continue to have strong growth. The figure below shows Hyperion Research’s projection for accelerators in the HPC market. Accelerators initially had a slow HPC adoption curve but have been steadily gaining momentum with GPUs emerging as the most popular option. Today, GPUs have become nearly synonymous with accelerators in the HPC space, given that GPUs comprise the overwhelming majority of accelerators in HPC clusters worldwide. Although originally designed as gaming processors for graphics rendering, GPUs were found to be well suited to a wide range of HPC applications and critical underlying AI functions, such as matrix multiplication, driving impressive performance gains. As a result, many HPC sites across a broad range of verticals are now using accelerators to speed up a larger portion of their increasingly AI-based workloads