- author = "Aymeric Blot and Justyna Petke",
- title = "{MAGPIE}: Machine Automated General Performance Improvement via Evolution of Software",
- howpublished = "arXiv",
- year = "2022",
- month = "4 " # aug,
- keywords = "genetic algorithms, genetic programming, genetic improvement, SBSE, parameter tuning, algorithm configuration, compiler optimisation, local search, multi-objective, running time, solution quality, metaheuristic-based search strategies, srcML, linear GP, GCC, LLVM, Clang, OpenJDK, GraalVM Java",
-
URL = "
https://arxiv.org/abs/2208.02811",
-
DOI = "
doi:10.48550/arxiv.2208.02811",
-
code_url = "https://github.com/bloa/magpie",
- size = "19 pages",
-
abstract = "Performance is one of the most important qualities of
software. Several techniques have thus been proposed to
improve it, such as program transformations,
optimisation of software parameters, or compiler flags.
Many automated software improvement approaches use
similar search strategies to explore the space of
possible improvements, yet available tooling only
focuses on one approach at a time. This makes
comparisons and exploration of interactions of the
various types of improvement impractical.
We propose MAGPIE, a unified software improvement framework. It provides a common edit sequence based representation that isolates the search process from the specific improvement technique, enabling a much simplified synergistic workflow. We provide a case study using a basic local search to compare compiler optimisation, algorithm configuration, and genetic improvement. We chose running time as our efficiency measure and evaluated our approach on four real-world software, written in C, C++, and Java.
Our results show that, used independently, all techniques find significant running time improvements: up to 25percent for compiler optimisation, 97percent for algorithm configuration, and 61percent for evolving source code using genetic improvement. We also show that up to 10percent further increase in performance can be obtained with partial combinations of the variants found by the different techniques. Furthermore, the common representation also enables simultaneous exploration of all techniques, providing a competitive alternative to using each technique individually.",
-
notes = "MiniSAT, LPG (v1.2 AClib), Sat4J, Weka (v3.8) random
forrest, Dexter dataset.
Written in Python 3.5+, MAGPIE is based on PyGGI 2.0
'Software is never done' 'Machine Automated General Performance Improvement through Evolution of software' 'Automated combination of the patches' (mutations) 'compiler and interpreter flag optimisation'
Figure 3: Cross-validation with k-folding and holdout.
p7 'only 10 training instances are actually used due to the very high evaluation cost'
StmtDelete, StmtReplace, StmtInsert, ConstantSet, ConstantUpdate, ParamSet.
fitness = CPU instructions (rather than CPU time).
GraalVM disabled use of 'compressed ordinary object pointers'.",
