Position: Professor Pronouns: she/her Office: TCL 309 Phone: 413.597.2772 E-mail: [email protected] Website |
Education
- Ph.D. Stanford University 2005
- M.S. Stanford University 2001
- B.S. Duke University 1997
Interests
- Parallel architectures, including heterogeneous architectures
- Internet of Things, particularly with respect to data consistency correctness
- Memory systems
- Workload characterization of emerging applications
Biography
Kelly received her Ph.D. from Stanford University in 2007. Her graduate work focused on how to distribute computation and data across chip multiprocessors in order to take advantage of available hardware resources while reducing communication distance on-chip. She has since worked on projects related to the characterization of emerging application domains and on ways to effectively improve the use of graphics processors with respect to power and performance. More recently, she has begun exploring the correctness of Internet of Things platforms and applications with respect to data consistency.
Research
Kelly is currently working on ways to verify the correctness of Internet of Things platforms and applications, with respect to data consistency. Additionally, she continues to explore ways to improve the performance and power of parallel computer architectures, including heterogeneous architectures. This work includes proposing software and hardware optimizations as well as developing tools that help programmers understand when specific optimizations or hardware are applicable.