Academic
Publications
The road not taken: Estimating path execution frequency statically
The road not taken: Estimating path execution frequency statically   (Citations: 7)
BibTex | RIS | RefWorks Download
A variety of compilers, static analyses, and testing frameworks rely heavily on path frequency information. Uses for such information range from optimizing transfor- mations to bug finding. Path frequencies are typically ob- tained through profiling, but that approach is severely re- stricted: it requires running programs in an indicative envi- ronment, and on indicative test inputs. We present a descriptive statistical model of path fre- quency based on features that can be readily obtained from a program's source code. Our model is over 90% accu- rate with respect to several benchmarks, and is sufficient for selecting the 5% of paths that account for over half of a program's total runtime. We demonstrate our technique's robustness by measuring its performance as a static branch predictor, finding it to be more accurate than previous ap- proaches on average. Finally, our qualitative analysis of the model provides insight into which source-level features indicate "hot paths."
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
Order by: