Jose Nelson Amaral is a professor in the Department of Computing Science. He is the head of the Compiler Design and Optimization Laboratory. His areas of expertise also include high-performance computing and computer architecture. He has published extensively in venues with strict peer-reviewing in these areas (see http://www.cs.ualberta.ca/%7Eamaral/epublications.html).

Amaral has active collaborations with IBM in the area of compiler design and implementation. Several of the ideas generated in Amaral's research group have been integrated into IBM commercial compilers. Amaral's research team was  the winner of the "Team of the Year" award in 2006 from the IBM Center for Advanced Studies in Toronto, Canada.

Prof. Amaral's extensive service to the scientific community includes membership in the program committee for many international conferences: proceedings chair for the International Parallel and Distributed Processing Symposium (for two years) and for the International Conference on High Performance Computing; Co-chair for the Workshop on Compiler-Driven Performance (four years); and external reviewer for many research proposals, tenure, and promotion cases.

Prof. Amaral is an associate editor for the IEEE Transactions on Computers. He is a Faculty Fellow with the IBM Center for Advanced Studies in Toronto and he has received the prestigious and highly competitive IBM Faculty Award for two years in a row.

• Compilers That Adapt After Leaving the Compiler Shop:

  Optimizing compilers often need feedback from training runs of programs to determine the frequency of execution of different paths and the affinity between data elements. Currently static compilers rely on static feedback data, ie. data that is ...

• Compilers that Fix Performance Bugs Generated by Good Programmers:

  Good programming and software engineering practices encourage the semantically-meaningful encapsulation of data into objects or data structures. These practices also encourage the design and implementation of relatively small, self contained, mo...

• Programming Systems that Increase Parallel Programmers' Productivity:

  A major issue that prevents the widespread adoption of parallel programming is the low productivity of parallel programmers. Typically writing a correct and efficient parallel program requires a more skilled programmer than writing the corr...