ABC: The Way It Should Have Been Designed

Research Scientist
Department of Electrical Engineering and Computer Science
University of California, Berkeley
Berkeley, California, USA

 

Webpage

Thursday, 28 September 2017 at 11:00 in room BC 420

Abstract:

Twelve years ago, in September 2005, the first public version of ABC was released.  It featured technology-independent synthesis by DAG-aware rewriting, technology mapping for standard cells and lookup tables, and simple combinational equivalence checking, all based on the And-Inverter Graphs (AIG) data-structure used to unify the computation flow.  In the coming years ABC has been adopted as an optimization engine and a research environment by a number of academic and industrial users.  The use that followed exposed a number of shortcomings in the original design of ABC.  This talk focuses on what is present and, more importantly, what is missing in ABC, and how ABC could be redesigned to make it more versatile and user-friendly. The motivation for this talk is to help academic researchers maximize the usefulness of their tools and set a new standard for future versions of ABC.

About the speaker:

Alan Mishchenko graduated from Moscow Institute of Physics and Technology (Moscow, Russia) in 1993 with M.S.and received his Ph.D. from the Glushkov Institute of Cybernetics (Kiev, Ukraine) in 1997. From 1998 to 2002 he was a visiting scientist at Portland State University in Oregon. In 2002, he joined the EECS Department at UC Berkeley, where he is currently a full research engineer. Alan's research interests are in developing computationally efficient methods for synthesis and verification.