On December 7, 2017 at 10 am PT (1 pm ET), Alliance member company Mentor (a Siemens Business) will deliver the free webinar "Optimizing Machine Learning Applications for Parallel Hardware," presented by Randy Allen, Director of Advanced Research in the company's Embedded Systems Division. From the event page:
Machine learning applications have a voracious appetite for compute cycles, consuming as much compute power as they can possibly scrounge up. As a result, they are invariably run on parallel hardware – often parallel heterogeneous hardware—which creates development challenges of its own. They are quite often developed by a theoretician who views the world from an abstract, Matlab-level view, but must be implemented and optimized by programmers whose view of the world is constrained by real time considerations. Squeezed between the egos of theoreticians, real time constraints, and delivery schedules, these programmers face a daunting task. In this webinar, Dr. Allen brings his decades of experience developing high performance compilers for parallel and fixed-point architectures to the aid of the machine learning programmer. Successfully optimizing a machine learning application requires an understanding of algorithms, architectures, and compilers, and Dr. Allen presents development methodologies and optimization tips garnered through his experience as a compiler-writer to help navigate programmers through the perils of real time constraints, hard delivery schedules, and the egos of theoreticians.
For more information and to register, please see the event page.
