Steve Blackburn

Steve co-created MMTk and leads both its development and the DaCapo benchmark project, providing shared infrastructure for language implementers and performance researchers. The 2025 work-packets paper demonstrates how eight garbage collectors can share common building blocks while supporting high performance. These contributions help engineers investigate performance and build memory management software that is easier to maintain, optimise and check.

Steve Blackburn is a Research Scientist at Google DeepMind and Professor of Computer Science at the Australian National University, focusing on programming language implementation and performance analysis. As a co-creator of MMTk, a reusable memory management toolkit, Steve helped establish shared infrastructure for language developers and researchers. The original Java implementation became part of the Jikes RVM research system in 2002, and a later rewrite led at the Australian National University made the toolkit usable across multiple language runtimes.

Steve also leads the DaCapo benchmark project, providing workloads for measuring software performance. A 2025 paper with Wenyu Zhao and Kathryn McKinley explores work packets: small groups of memory management tasks with explicit processing and scheduling rules. Evaluated across eight garbage collectors, the design combines reusable code with high performance and simpler implementation. Academic and community leadership includes conference steering and service as Associate Dean at the Australian National University from 2014 to 2019, with responsibility for Diversity and Inclusion from 2016 to 2019.