Kunle Olukotun : 2013 Plenary Session

Big data analytics applications present many opportunities for high performance computation on heterogeneous systems that are composed of shared memory multiprocessor nodes, GPUs and distributed memory clusters. Each of these architectures has its own architecture-specific programming model, which makes it difficult to develop a single big data application that runs efficiently on all of these architectures. Domain-specific languages (DSLs) and compilers provide a solution to this problem by providing the capability for high-level application-specific abstractions to be mapped directly to low-level architecture-specific programming models resulting in both high programmer productivity and high execution performance. In this talk, I will describe a set of DSLs for data analytics, including DSLs for data extraction and transformation, data querying, graph analysis, and machine learning. Using a system called Delite, these DSLs can be composed and optimized together to produce efficient execution on heterogeneous systems.

Kunle Olukotun is a Professor of Electrical Engineering and Computer Science at Stanford University. Olukotun is best known as a pioneer in chip multiprocessor (CMP) design and the leader of the Stanford Hydra CMP research project. Olukotun founded Afara Websystems to develop high-throughput, low-power server systems with CMP technology. The Afara microprocessor, called Niagara, was acquired by Sun Microsystems. Niagara derived processors now power all Oracle SPARC-based servers. Olukotun currently directs the Stanford Pervasive Parallelism Lab (PPL), which seeks to proliferate the use of heterogeneous parallelism in all application areas using Domain Specific Languages (DSLs). Olukotun is an ACM Fellow and IEEE Fellow.