Abstract
A “linearized” scheme of polyvariant specialization for imperative languages is described in the paper. The scheme is intended for increasing efficiency of specialization. Main properties of the scheme are linear generation of residual code and single memory shared by different variants of specialization process. We sketch a new kind of analysis, called configuration analysis, which is aimed on reducing memory usage during specialization. To demonstrate the advantages of the proposed scheme we discuss some benchmarks for the MPMix partial evaluator for Modula-2 language.
Cite
CITATION STYLE
Bulyonkov, M. A., & Kochetov, D. V. (1996). Practical aspects of specialization of algol-like programs. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1110, pp. 17–32). Springer Verlag. https://doi.org/10.1007/3-540-61580-6_2
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
