Antonino Tumeo
Chief Scientist – High Performance Computing – Pacific Northwest National Laboratory
Invited Talk: SODA: An End-To-End Open-Source Hardware Compiler for Machine Learning Accelerators
Slides
Video
Abstract:
Enabling autonomous control in novel scientific experimental workflows requires the ability to generate highly specialized systems for data analysis and artificial intelligence, enabling the low-latency reasoning capabilities needed to take real-time decisions. This paper presents the SODA (Software Defined Accelerators) framework, an open-source modular, multi-level, no-human-in-the-loop, hardware compiler that enables end-to-end generation of specialized accelerators from high-level data science frameworks. SODA is composed of SODA-Opt, a high-level frontend developed in MLIR that interfaces with domain-specific programming environments and allows performing system level design, and Bambu, a state-of-the-art high-level synthesis (HLS) engine that can target different device technologies. The framework implements design space exploration as compiler optimization passes. We show how the modular, yet tight, integration of the high-level optimizer and lower-level HLS tools enables the generation of accelerators optimized for the computational patterns of converged applications.
Biography:
Dr. Antonino Tumeo received the M.S degree in Informatic Engineering, in 2005, and the Ph.D degree in Computer Engineering, in 2009, from Politecnico di Milano in Italy. Since February 2011, he has been a research scientist in the PNNL’s High Performance Computing group. He Joined PNNL in 2009 as a post doctoral research associate. Previously, he was a post doctoral researcher at Politecnico di Milano. His research interests are modeling and simulation of high performance architectures, hardware-software codesign, FPGA prototyping and GPGPU computing.