Jeronimo Castrillon
Dresden University of Technology
Models and Languages to Tame Heterogeneous Computing Systems
Programming heterogeneous computing systems remains a daunting task, especially in the presence of emerging architectures and stringent application constraints. This talk discusses the importance of formal models of computation and domain-specific languages (DSLs), as a complement to parallelising compilers, high-level synthesis tools and mainstream parallel programming models.
We argue that formal semantics are needed to provide better guarantees, enable coarser-grained optimization and improve overall productivity (coding and debugging). Concretely, the talk presents recent results on compiler-guided adaptive execution of dataflow programs, time-deterministic execution of dataflow models extended with discrete event semantics, and higher-level DSLs and optimizing compilers for tensor expressions with an outlook for emerging technologies.
His research interests lie on methodologies, languages, tools and algorithms for programming complex computing systems. He has more than 100 international publications and has been a member of technical program and organization committees in international conferences and workshops (e.g., DAC, DATE, ESWEEK, CGO, LCTES, Computing Frontiers, ICCS and FPL). He is also a regular reviewer for ACM and IEEE journals (e.g., IEEE TCAD, IEEE TPDS, ACM TODAES and ACM TECS). In 2014, Prof. Castrillon co-founded Silexica GmbH/Inc, a company that provides programming tools for embedded multicore architectures. Prof. Castrillon is Senior Member IEEE, Member ACM and was a founding member (2017-2019) of the executive committee of the ACM “Future of Computing Academy” (FCA)
