Abstract
This paper describes a program profiling and analysis tool called Gleipnir. Gleipnir collects memory access traces and associates each access with a specific program internal structure such as a thread, a function, a data structure or a scalar variable. The data provided by Gleipnir can be used to analyze how program variables and associated memory accesses map to L-1 as well as higher level cache memories. This information can be used to investigate techniques to refactor data or code to improve memory access performance. It is our hypothesis that optimizing cache performance at all levels is very important to both single-core and multi-core processors. In this paper we will describe the Gleipnir tool and some examples of its use in optimizing memory performance. The overall goal of our research is to develop techniques usable by application developers, compilers, and runtime systems to improve the performance of their applications.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
