Abstract
Traditional performance and energy scaling benefits based on technology improvements have slowed greatly. At the same time the demand for computing capability is unsatiated with new killer applications emerging in the domains of robotics, automotive, and machine-intelligence. Lack of progress in technology scaling will necessarily place more demands on the computer architecture and software layers to deliver capability. This talk outlines two promising directions for research in this coming Golden Age of design. First, hardware acceleration in the form of data path and control circuitry customized to particular algorithms or applications has surfaced as a promising approach, as it delivers orders of magnitude performance and energy benefits compared to general-purpose solutions. To broaden the scope of applicability for accelerator-based architectures it will be necessary to preserve flexibility and provide programmable solutions while reducing design costs. At the same time, the last decade has seen great progress in compiler and architecture approaches to exploit parallelization. Continued research investment in this direction will lead to solutions with better scalability and even more radical approaches are on the horizon. This talk will draw highlights from several research projects in both directions.