Jeanne Ferrante

Associate Dean of the Jacobs School
Associate Vice Chancellor for Faculty Equity
Professor, Computer Science & Engineering

Compiling techniques that make scientific applications run faster.

In compiling, applications are translated from high-level programming languages such as C and Fortran and their variants (among others) to machine-executable form. Professor Ferrante focuses on a middle stage of compiling process: performance optimization. She also has worked on program optimizations to increase parallel processing and is starting to participate in research to develop "application signatures" to help predict the best architectures for a given application. A current focus is how to best identify code sections likely to gain most from predicated execution, the capability that distinguishes Intel's new 64-bit Itanium architecture (formerly IA-64). She has led research with Professor Larry Carter into "hierarchical tiling," in which data is stored in chunks or "tiles" convenient to transfer through the various data-storage levels. Ferrante's research into performance was recently extended to grid computing, or networks of computers that tie together various types of machines, some of them separated by great distances. She is one of several UCSD professors working on bandwidth-centric computing, or the development of algorithms that focus on maximizing steady-state throughput when scheduling work across a grid.

Capsule Bio:
Jeanne Ferrante joined the UCSD faculty in 1994. She is one of three professors in the High Performance Compilers group, and was vice chair of Computer Science & Engineering in the 1999-2000 academic year. From 1978-1994, she was a research staff member at IBM's T.J. Watson Research Center. Among other honors, she was selected ACM fellow in 1996. She received her Ph.D. in mathematics from MIT in 1974.

Institute Affiliation:
Senior Fellow,
San Diego Supercomputer Center »