Valerie Zhao ’18 and Ben Wood participated in SPLASH 2017, the ACM SIGPLAN Conference on Systems, Programming Languages, Applications: Software for Humanity, in Vancouver, BC, in late October. SPLASH is an umbrella for several conferences and workshops in the area of programming languages.

Valerie won 3rd prize in the undergraduate category of the Student Research Competition, where she presented a poster and talk on her summer research work, Abstracting Resource Effects, undertaken at Carnegie Mellon University with Darya Melicher, Jonathan Aldrich, and Alex Potanin. Valerie’s collaborator Darya Melicher, a PhD student at CMU, presented more of their work at the OCAP workshop. Their work introduces a novel effect system that supports rigorous checking of how programs use system resources in a security-focused programming language.

Ben gave a talk on his OOPSLA paper, Instrumentation Bias for Dynamic Data Race Detection, with collaborators from Google, The Ohio State University, and the University of Washington. Their work introduced a software system for accurately detecting data races, a problematic type of concurrent programming error. The analysis helps eliminate a source of performance overhead in error detection by exploiting properties of common program patterns.

Earlier in October, Ben’s collaborator gave a talk on their paper, PARSNIP: Performant Architecture for Race Safety with No Impact on Precision, at the 50th ACM/IEEE Symposium on Microarchitecture (MICRO), in Cambridge.  This work, with collaborators at U. Penn, designed efficient hardware support for data race detection that, combined with software techniques, could eventually provide always-on concurrency error detection much like modern memory-safe languages provide explicit runtime exceptions for null dereferences or array bounds errors.