High Performance Computing Technologies, Solutions to Exascale Systems, and Beyond
This course program, developed in partnership with IEEE Future Directions, starts by briefly exploring the historical evolution of "big iron" computers from the 70s, 80s, and 90s, emphasizing their relevance in today's exascale systems. It delves into the shift from optimization concerns in the 90s to structuring applications for optimal hardware utilization. Then focuses on high-performance computing, and how to address challenges and solutions in the Exascale era. The role of HPC as a platform for scientific discovery is examined, with real-world applications such as rapid-response drug discovery during the COVID-19 pandemic. It showcases how the increasing demand for advanced models using enormous amounts of data is met through hardware and software optimization. The current and future CPU and GPU contexts, programming languages in HPC, parallel programming models, and workload management are also discussed. Last, the use of AI in scientific research and emerging technologies for improved AI delivery in HPC applications is extensively discussed with real life examples.
What you will learn:
- Understand how supercomputer architectures from the 60s, 70s, & 80s, on-processor parallelism, the decreasing clock cycle in the late 90s, and the introduction of GPUs and multi-core nodes affect high performance computing today
- Describe high performance computing evolution and what is the exascale era of computing. How exascale enables the discovery of actionable insights based on real-life examples or “grand challenge” problems as happened with the COVID-19 drug/vaccine research
- Understand how it is possible to accelerate application performance via choice of processor, the memory hierarchy, the interconnect, the architecture, and the choice of storage hardware
- Describe the expected hardware components and systems that will be prevalent in the near future in HPC systems. Additionally, the software used to program and run applications in modern systems will be discussed
- To learn about the leading edge of HPC research that involves understanding how AI can be used in science in addition to the latest developments in heterogeneity. To see examples of serverless computing that portray the delivery of HPC as a service. To describe those technologies that can be used to better deliver AI for science and HPC
Courses included in this program:
Course Program Length: 5 hours
Who Should Attend: National grid engineers and managers, microgrid engineers, minigrid engineers, minigrid project managers, minigrid developers, microgrid developers, minigrid entrepreneurs, policy and regulatory authorities
Instructors
John Levesque, Hewlett Packard Mission Critical Systems Chief Technology Office
Sreenivas Rangan Sukumar, Hewlett Packard Mission Critical Systems Chief Technology Office
Dejan Milojicic, Distinguished technologist, Hewlett Packard Labs
Paolo Faraboschi, Vice President and HPE Fellow and directs Artificial Intelligence Research at Hewlett Packard Labs
Barbara Chapman, Distinguished Technologist for the Cray Programming Environment at Hewlett Packard Enterprise (HPE)
Larry Kaplan, Senior Distinguished Technologist at HPE
Publication Year: 2023
ISBN: 978-1-7281-7807-3