Effective Static Race Detection

Date

January 17, 2006

Speaker

Alex Aiken

Affiliation

Stanford University

Overview

Despite years of research on static techniques, the state of practice for analyzing concurrent software remains relatively primitive. Recent advances in program analysis technology, however, hold out some hope for significant progress. This talk will present the approach and initial results of a project to build an effective static race detector, namely one that developers can use routinely to reliably identify dangerous races in realistic scale systems. We’ll discuss how far we’ve advanced toward that goal as well as what problems remain. (Joint work with Mayur Naik.)

Speakers

Alex Aiken

Alex Aiken is a Professor of Computer Science at Stanford University. Alex was previously a Research Staff Member at the IBM Almaden Research Center http://www.almaden.ibm.com/ and a Professor in the EECS department at UC Berkeley http://www.eecs.berkeley.edu before joining the Stanford faculty in 2003.

‚Äč