Compiler technologies can improve the performance of programs on existing hardware; however, often require expert knowledge from the user to deliver the best performance. Techniques like domain-specific languages, auto-parallelization, and auto-vectorization can lower the barrier for high performance for non-experts.
Compiler technologies can improve the performance of programs on existing hardware; however, often require expert knowledge from the user to deliver the best performance. My research focuses on lowering the barrier for non-experts to unlock better performance from their programs. I am interested in developing frameworks for building domain-specific languages, and improving compilers' ability to auto parallelize and vectorize programs.
Justin Szaday
Profile Photo
Previous degree(s)
BS, Computer Engineering, Valparaiso University
PhD start year
2016
Award Year
2016
Department/Unit
Computer Science
Research Advisor(s)
David Padua
Description of Research
Email
szaday2@illinois.edu
Research Area
Compilers