The next generation of high-performance computers (HPC) has arrived, which will provide researchers and scientists with powerful latest tools for support in acceleration of scientific discoveries and driving innovations. Intel has always been at the forefront of this convergence of Artificial Intelligence (AI), simulation, analytics, modeling and other such high-performance computing workloads which can move the industry towards a new era in supercomputing.
Intel was again seen as a leader in HPC at ISC 2018 with today’s majority of supercomputing platforms relying on Intel Xeon processors as the most preferred processor in the globe’s leading 500 supercomputers. These processors have the remarkable power of providing flexibility and performance for handling the most difficult HPC workloads at any scale that any industry or science may require.
Staying true to its tradition, Intel continues to innovate and deliver high-speed interconnected technologies which enable the deployment of cost-effective HPC systems. At ISC 2018, Intel reportedly shared a next-generation Omni-Path Architecture (Intel OPA200), which is scheduled to be released in 2019. It is all set to provide the data rate speed up to 200 GB per second which is double to the performance of the previous generation of computers. This latest generation of computing is compatible and interoperable with the present generation of Intel OPA. This new generation of computers also possesses high-performance capabilities and will enable the system architects to scale tens of thousands of nodes while simultaneously benefiting from the improved combined ownership cost.
Intel also announced another breakthrough in computing technology at ISC 2018 namely the Intel Select Solution for Professional Visualization. It is a conveniently deployed, Intel-optimized reference architecture system, purposely-built to fulfill the demands of the present era’s most complicated challenges related to data explosion. Its functioning is based on the usage of the platform’s onboard memory for graphical rendering of huge datasets in real time. As there is a larger footprint of memory available in these Intel solutions, they are best suited for bigger data sets inside HPC workloads as compared to computers with captive memory pools.
It would be interesting to see what more Intel has to offer to the super-computing industry in the near future.