Intelligent Software for Scientific Computing

Warren Armstrong (School of Computer Science, CECS ANU)

CS HDR MONITORING CompSys Research Group

DATE: 2010-04-08
TIME: 09:00:00 - 09:30:00
LOCATION: Ian Ross Seminar Room
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
Maintaining efficiency in long-running programs often requires adapting to changes in how the program interacts with its computational environment. These changes can come as a natural consequence of progressing a computation, as a result of competition for hardware resources, or through the migration of the program between heterogeneous cluster nodes. Intelligent software is capable of monitoring its execution and adapting itself to suite changing circumstances.

I am investigating intelligent software in the domain of sparse matrix-vector products. I will present simulation data showing the potential for faster computation when sparse matrix storage formats are selected at runtime. I will report on microbenchmarks that measure DynInst, a third party library that performs binary modification of processes. I will finish by outlining plans to use DynInst to implement runtime sparse matrix format selection.

BIO:
PhD student in Computer Science. https://cecs.anu.edu.au/user/3908