Wednesday, April 19th, 2017 at 2:00 pm in Rice 128 (Young Board Room)
Committee: Kevin Skadron (Advisor), Samira Khan (Chair), Andrew Grimshaw, Gabriel Robins and Mircea Stan.
Title: Accelerating Pattern Recognition Processing with Hybrid von Neumann/Spatial Architectures
Newly available spatial architectures to accelerate finite automata processing for pattern recognition have spurred a large amount of research and development of finite automata-based applications and accelerator research. However, a lack of standard, open-source tools for automata processing application and architecture research has slowed the pace of innovation.
This proposal outlines five new contributions to the area of automata processing: 1) a new non-obvious use case for automata processing: efficient and high-quality pseudo-random number generation, 2) a novel automata processing simulation, profiling, and optimization framework to accelerate automata application and architecture research, 3) a diverse benchmark suite of standardized automata graphs and inputs from published work to provide easy and fair comparisons among automata processing engines and architectures, and 4) a toolchain to enable an entire field of research: spatial, reconfigurable automata processing architecture research. We present the capabilities of these tools as well as results from investigations using these tools, highlighting their usefulness and potential.
Motivated by preliminary results from the above tools and investigations, this proposal recognizes the potential for hybrid von Neumann/spatial automata processing acceleration. Based on these results we propose the investigation of 5) a hybrid von Neumann/spatial automata processing accelerator.