RT conference output
T1 Automatic Software Performance Optimization using Genetic Algorithms
A1 Aragón Jurado, José Miguel
A1 Torre Macías, Juan Carlos de la
A1 Ruiz Villalobos, Patricia
A1 Dorronsoro Díaz, Bernabé
A2 Ingeniería Informática
AB Software programs (SW) are who ultimately drives the underlying hardware (HW). Therefore SWshould make an efficient use of all available resources in order to achieve high performance levelswhen executing.Compilers typically perform a number of code transformations, also known as optimizations,before generating the executable file. The purpose of applying such optimizations is to modify theuse of the underlying HW by the considered SW. These are generic sequences of optimizationsthat are known to generally perform well. However, it is clear that distinct SW programs and HWarchitectures offer different characteristics, and the same sequence of code transformations cannotfully exploit them for optimal performance.Therefore, there is a need of finding ad hoc optimization sequences to accurately optimize theperformance of every SW/HW pair. However, this is a titanic task, taking into account the immensevariety of existing HW architectures, for which every SW must be optimized. In addition, theprocess of optimizing SW performance for a given HW architecture is a difficult, error prone, andexpensive task, requiring the hands of expert programmers with high knowledge on the consideredarchitecture. Automatizing this process would allow the generation of a new generation of smartcompilers, able to perform ad hoc transformations for the considered SW and HW for optimalperformance.We propose in this work a novel approach to automatize the process of optimizing the performanceof a particular SW program on a specific HW. For that, we define a novel combinatorialoptimization problem, that we call SCOP (for Software Code Optimization Problem), consistingin finding a sequence of code transformations that minimizes the runtime of the resulting SW onthe considered HW architecture. This extended abstract presents our work [2], recently publishedin IEEE Transactions on Sustainable Computing.
YR 2024
FD 2024
LK http://hdl.handle.net/10498/35880
UL http://hdl.handle.net/10498/35880
LA eng
DS Repositorio Institucional de la Universidad de Cádiz
RD 11-nov-2026