Abstract
The term "instrumentation" refers to modification of a program or its runtime environment to make hidden details of execution visible. Instrumentation can severely compromise program execution speed. Frameworks like DTrace (Sun Microsystems) and VProbes (VMware) offer practical ways of addressing performance concerns, but there has been no formal understanding of what it means for instrumentation to be efficient. To fill this gap, we propose a criterion based on that of Popek and Goldberg for virtual machines and on our previous work relating this to Jones optimality of program specializers. We further suggest linguistic constraints on instrumentation code to enable more aggressive static optimization of dynamically instrumented programs. © 2010 Springer-Verlag.
Cite
CITATION STYLE
Feigin, B., & Mycroft, A. (2010). Formally efficient program instrumentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6418 LNCS, pp. 245–252). https://doi.org/10.1007/978-3-642-16612-9_19
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
