Hybrid partial evaluation

Author:

Shali Amin1,Cook William R.1

Affiliation:

1. The University of Texas at Austin, Austin, TX, USA

Abstract

Hybrid partial evaluation (HPE) is a pragmatic approach to partial evaluation that borrows ideas from both online and offline partial evaluation. HPE performs offline-style specialization using an online approach without static binding time analysis. The goal of HPE is to provide a practical and predictable level of optimization for programmers, with an implementation strategy that fits well within existing compilers or interpreters. HPE requires the programmer to specify where partial evaluation should be applied. It provides no termination guarantee and reports errors in situations that violate simple binding time rules, or have incorrect use of side effects in compile-time code. We formalize HPE for a small imperative object-oriented language and describe Civet, a straightforward implementation of HPE as a relatively simple extension of a Java compiler. Code optimized by Civet performs as well as the output of a state-of-the-art offline partial evaluator.

Publisher

Association for Computing Machinery (ACM)

Subject

Computer Graphics and Computer-Aided Design,Software

Cited by 9 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Partial evaluation of string obfuscations for Java malware detection;Formal Aspects of Computing;2017-01

2. The Essence of Multi-stage Evaluation in LMS;A List of Successes That Can Change the World;2016

3. Zero-overhead metaprogramming: reflection and metaobject protocols fast and without compromises;ACM SIGPLAN Notices;2015-08-07

4. Partial Evaluation for Java Malware Detection;Logic-Based Program Synthesis and Transformation;2015

5. A Refactoring Library for Scala Compiler Extensions;Lecture Notes in Computer Science;2015

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3