Compactly Numbering the Approximately Most Executing Path for k-iteration
Priya Arundhati, Santosh Kumar Pani · 2020 Fourth International Conference on Inventive Systems and Control (ICISC) · 2020
In the present day scenario, the computation is demanding reduced energy with some tolerable loss in quality. Approximate computing provides this feature. The approximation can be applied in various levels of the execution stack. One of the levels is a software approximation technique. This can be applied in various software engineering applications and program analysis techniques like in path profiling which can be applied in various fields. In this paper, the concept of randomized loop perforation is applied to the existing preferential path profiling algorithm for k-iteration. The proposed algorithm produces the compact number for the path identifier of the AME (Approximately Most Executing) paths which reduces the overhead of Path profiling by the use of Array instead of hash table implementation.