This talk showcases a methodology with examples to break down applications to low-level primitives and identify optimizations on existing compute instances or platform or for offloading specific portions of the application to accelerators or GPU’s. With the increasing use of a combination of CPU, GPU and accelerators/ASIC’s, this methodology could prove increasingly useful to evaluate what kind of compute to use and when.
Ravi is a Sr. Principal Engineer in Intel's Datacenter and AI group leading software engineering collaboration with hyperscale cloud providers. He focuses on performance engineering to address performance bottlenecks, infrastructure overhead and efficiency. Prior to this, he was responsible for the design and implementation of Operating Systems, Configuration Management and Compute Metrics solutions for Intel's internal design computing environment which runs on millions of cores of compute. He is a Lean Six Sigma Green Belt and has approved patent filings related to Energy Efficiency, Distributed Computing and Check-pointing. Ravi has authored numerous whitepapers including IEEE publications, IT@Intel public whitepapers and internal conference papers.
Harshad Sane is a performance engineer in the Data Center and AI group with a deep technical expertise in system software, memory, and CPU architectures. He specializes in performance monitoring, software optimization, and tool development with focus in the cloud domain. Harshad joined Intel in 2008 after completing his undergraduate degree in Electrical and Telecommunications from College of Engineering and Technology (COET), Pune, India and M.S. in Electrical and Computer Engineering from UC Boulder.
