Nvidia, Intel chips continue to dominate the HPC landscape

(Image credit: Shutterstock / Timofeev Vladimir)

The majority of world's supercomputers continue to run on Intel processors and make use of Nvidia GPU accelerators, according to the latest bi-annual Top500 rankings.

The rankings reveal that, with a share of 86.2%, Intel continues to be the dominant processor provider of the top 500 systems, with Skylake accounting for 35.2%, followed by Broadwell at 24.4%, and Cascade Lake at 16.8%.

Furthermore, a total of 145 systems on the list are using accelerator/co-processor technology, with 96 of these relying on Nvidia Volta chips, while 26 use Nvidia Ampere, and nine run Nvidia Pascal.

More of the same

There isn’t much change in the list of top ten supercomputers, with Fugaku, powered by Fujitsu’s custom Arm A64FX processor, retaining the top spot with three-times the performance of IBM Summit at number two.

The top ten does get a new entrant in the form of the Perlmutter supercomputer designed by China’s National Research Center of Parallel Computer Engineering & Technology (NRCPC).

Installed at the Department of Energy at the Lawrence Berkeley National Laboratory in the US, Perlmutter is based on the HPE Cray “Shasta” platform. The supercomputer is a heterogeneous system with both Nvidia A100-powered GPU-accelerated and AMD EPYC-based CPU-only nodes.

Also of note is the fact that, while Intel’s 86.2% share of the top 500 has come down from 91.8% six months ago, AMD has increased its share from 4.2% to 9.8% in the same period.

Overall, AMD powers a total of 49 supercomputers, of which 43 use AMD Rome. Also, while three of the top 10 supercomputers use AMD, the majority of the Green 500 energy-efficient supercomputers are powered by the company's chips.

Here's our list of the best cloud storage services right now

TOPICS

With almost two decades of writing and reporting on Linux, Mayank Sharma would like everyone to think he’s TechRadar Pro’s expert on the topic. Of course, he’s just as interested in other computing topics, particularly cybersecurity, cloud, containers, and coding.