Colloquium | September 7 | 4-5 p.m. | Soda Hall, 306 (HP Auditorium)
David Patterson, Distinguished Engineer, Google
Electrical Engineering and Computer Sciences (EECS)
The success of deep neural networks (DNNs) from Machine Learning (ML) has inspired domain specific architectures (DSAs) for them. Googles first generation DSA offered 50x improvement over conventional architectures for ML inference in 2015. Google next built the first production DSA supercomputer for the much harder problem of training. Subsequent generations greatly improved performance of both phases. We start with ten lessons learned from such efforts.
The rapid growth of DNNs rightfully raised concerns about their carbon footprint. The second part of the talk identifies the 4Ms (Model, Machine, Mechanization, Map) that, if optimized, can reduce ML training energy by up to 100x and carbon emissions up to 1000x. By improving the 4Ms, ML held steady at
CA, jennyj@berkeley.edu, 5106427699
Jenny Jones, jennyj@berkeley.edu, 510-642-7699
Coming Soon, available September 7, 2022: David Patterson: A Decade of Machine Learning Accelerators: Lessons Learned and Carbon Footprint