Intel Expands the Boundaries of High-Performance Computing and Strengthens the Integration of High-Performance Computing and AI

The 2020 CCF National High Performance Computing Academic Annual Conference officially kicked off. At the conference, Intel shared in depth how to promote the further integration of high-performance computing and artificial intelligence, and demonstrated the industry vision of building a fusion future through the XPU strategy, thereby creating a more flexible computing architecture for customers with diverse computing needs in the data-centric computing era, and providing leading performance for various workloads.

Trish A. Damkroger, Vice President of Data Platforms Group and General Manager of High Performance Computing at Intel, delivered a speech at the conference

Trish A. Damkroger, vice president of Intel's Data Platforms Group and general manager of high-performance computing, said in her speech at the conference: "The outbreak of the COVID-19 pandemic this year has made us more clearly aware of the important role that high-performance computing plays in today's society. With the continuous integration of high-performance computing and artificial intelligence workloads, Intel is focusing on integrating edge to cloud, hardware and software into the entire ecosystem through the XPU strategy, leading high-performance computing into a new era of greater intelligence and promoting high-performance computing to play an important role in solving more major social challenges."

Today, the increasingly rich use cases of artificial intelligence are leading us into a new era of further enhanced computing intelligence. Intel realizes that the integration of high-performance computing and artificial intelligence will be an inflection point in the high-performance computing industry, and heterogeneity will be the key to accelerating the integration of high-performance computing and artificial intelligence. To this end, Intel will launch its first 10nm Ice Lake server CPU this year, and will launch the next-generation Xeon Scalable Processor Sapphire Rapids based on enhanced SuperFin technology in 2021 to more effectively support deep learning acceleration and the integration of high-performance computing and artificial intelligence workloads. In addition, Intel also plans to launch an independent general-purpose GPU "Ponte Vecchio" based on the Xe architecture. "Ponte Vecchio" is manufactured using a 7nm process and uses Intel Foveros 3D and Embedded Multi-Chip Interconnect Bridge (EMIB) innovative packaging technology to optimize high-performance computing and artificial intelligence workloads.

Intel has created unparalleled depth, breadth and performance in multiple product lines of high-performance computing, and has brought comprehensive acceleration and improvement to the field of high-performance computing under the guidance of six major technology pillars. Among them, OneAPI, launched by Intel in cooperation with the ecosystem, can simplify development across multiple types of processors and accelerators, and support a wide range of industry ecosystems to adopt this technology to promote innovation, simplify application development, and achieve scalability of applications on a large number of nodes; Intel's revolutionary memory and storage technology can eliminate storage bottlenecks and provide support for the next generation of workflows. In addition, Intel also provides strong support for the integration of high-performance computing and artificial intelligence workloads through intelligent interconnect technology, higher platform security, innovative process and packaging technology, and a variety of advanced architectures.

For a long time, in addition to continuously promoting technological innovation, Intel has also always focused on the training and development of high-performance computing talents.

Since 2013, Intel has sponsored the Intel Cup Parallel Application Challenge (PAC). The competition aims to popularize and cultivate students' parallel computing ideas and parallel computing system capabilities, and improve students' competitiveness in the global talent market. After years of development, the PAC has attracted more than 1,200 teams from more than 40 cities, more than 300 universities and more than 160 organizations in China, with a total of more than 7,000 participants.

During this year's 8th Intel Cup National Parallel Application Challenge (PAC2020), 12 teams from many well-known domestic universities such as the Chinese Academy of Sciences High Energy Physics Research Institute, the Chinese Academy of Sciences Computing Center, Zhejiang University, and the University of Science and Technology of China entered the finals. The entries covered multiple hot fields such as ocean, biomedicine, geology, astronomy, and AI deep learning. The challenge finals and award ceremony will be held during the 16th Annual Conference on High Performance Computing from September 28 to 30. In the future, Intel will continue to invest and pay attention to the cultivation of outstanding talents in the field of parallel computing, promote the updating and construction of parallel computing courses in Chinese universities, and will work with more Chinese production, education, and research institutions to promote the development of high-performance computing to a new level.