Automatic Staging via Partial Evaluation Techniques

13 pages•Published: March 27, 2016

Abstract

Partial evaluation and staging are two of the well-known symbolic
manipulation techniques of programs which generate efficient
specialized code. On one hand, partial evaluation provides us with an
automatic means to separate a program into two (or more) stages but
its behavior is perceived as hard to control. On the other hand,
staging (or staged calculus) requires us to manually separate a
program but with full control over its behavior. In the previous
work, the first author introduced a framework to relate the two
techniques, giving a unified view to the two techniques. In this
paper, we extend the framework to handle the cross-stage persistence
(CSP) and show that the 2-level staging annotation obtained by the
automatic separation is the best staging annotation in a system where
CSP is allowed for base-type values only. In the presence of CSP for
higher-type values, on the other hand, there is no single annotation
that is better than all the other annotations.

Keyphrases: binding-time analysis, cross-stage persistence, partial evaluation, staging

In: James H. Davenport and Fadoua Ghourabi (editors). SCSS 2016. 7th International Symposium on Symbolic Computation in Software Science, vol 39, pages 1--13

Links:	https://easychair.org/publications/paper/Prd
	https://doi.org/10.29007/rfbk