In this video from SC18, Dr. Eng Lim Goh from HPE describes the Convergence of HPC and Ai.
Data analytics and insights are fueling innovation across scientific research, product and service design, customer experience management, and process optimization. Real-time analytics with in-memory computing, big data analytics, insights from simulation and modeling fueled by HPC and predictive analytics with artificial intelligence (AI) are core capabilities required by data-driven organizations looking to gain competitive advantage with their digital transformation initiatives.
HPC enables complex modeling and simulation to accelerate innovation in diverse areas—ranging from molecular chemistry to genome sequencing, energy exploration, and financial trading. AI is the foundation for cognitive computing, an approach that enables machines to mimic the neural pathways of the human brain to analyze vast datasets, make decisions in real time, and even predict future outcomes. In order to succeed, all organizations need advanced technology solutions to support their HPC, Accelerated Analytics, and AI applications to execute increasingly difficult tasks and forecast evolving trends, equipping them to solve some of the world’s biggest scientific, engineering, and technological problems.
HPC is a driving force behind business growth and innovation, empowering users to execute compute- and data-intensive workloads quickly and accurately. Purpose-built HPC solutions are accelerating performance and increasing operational efficiencies like never before, which allows organizations to continuously scale and adopt cutting-edge tools in order to take on the next great challenge. It is also being recognized that HPC environments could be a good foundation for AI and deep learning, as they provide the extreme levels of scalability, performance, and efficiency required by these complex applications. With compute innovations such as these, IT departments can confidently implement and accelerate both HPC and AI applications and put their explosive data volumes to work.
