Access-based localization with bypassing

Hakjoo Oh, Kwangkeun Yi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

We present an extension of access-based localization technique to mitigate a substantial inefficiency in handling procedure calls. Recently, access-based localization was proposed as an effective way of tightly localizing abstract memories. However, it has a limitation in handling procedure calls: the localized input memory for a procedure contains not only memory locations accessed by the procedure but also those accessed by transitively called procedures. The weakness is especially exacerbated in the presence of recursive call cycles, which is common in analysis of realistic programs. In this paper, we present a technique, called bypassing, that mitigates the problem. Our technique localizes input memory states only with memory locations that the procedure directly accesses. Those parts not involved in analysis of the procedure are bypassed to transitively called procedures. In experiments with an industrial-strength global C static analyzer, the technique reduces the average analysis time by 42%. In particular, the technique is especially effective for programs that extensively use recursion: it saves analysis time by 77% on average.

Original languageEnglish
Title of host publicationProgramming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings
Pages50-65
Number of pages16
DOIs
Publication statusPublished - 2011 Dec 26
Externally publishedYes
Event9th Asian Symposium on Programming Languages and Systems, APLAS 2011 - Kenting, Taiwan, Province of China
Duration: 2011 Dec 52011 Dec 7

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7078 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference9th Asian Symposium on Programming Languages and Systems, APLAS 2011
CountryTaiwan, Province of China
CityKenting
Period11/12/511/12/7

Fingerprint

Data storage equipment
Recursion
Cycle
Experiments
Experiment

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Oh, H., & Yi, K. (2011). Access-based localization with bypassing. In Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings (pp. 50-65). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7078 LNCS). https://doi.org/10.1007/978-3-642-25318-8_7

Access-based localization with bypassing. / Oh, Hakjoo; Yi, Kwangkeun.

Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings. 2011. p. 50-65 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7078 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Oh, H & Yi, K 2011, Access-based localization with bypassing. in Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7078 LNCS, pp. 50-65, 9th Asian Symposium on Programming Languages and Systems, APLAS 2011, Kenting, Taiwan, Province of China, 11/12/5. https://doi.org/10.1007/978-3-642-25318-8_7
Oh H, Yi K. Access-based localization with bypassing. In Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings. 2011. p. 50-65. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-25318-8_7
Oh, Hakjoo ; Yi, Kwangkeun. / Access-based localization with bypassing. Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings. 2011. pp. 50-65 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{5c9ddf7122474f2593ddb4db13f0d2af,
title = "Access-based localization with bypassing",
abstract = "We present an extension of access-based localization technique to mitigate a substantial inefficiency in handling procedure calls. Recently, access-based localization was proposed as an effective way of tightly localizing abstract memories. However, it has a limitation in handling procedure calls: the localized input memory for a procedure contains not only memory locations accessed by the procedure but also those accessed by transitively called procedures. The weakness is especially exacerbated in the presence of recursive call cycles, which is common in analysis of realistic programs. In this paper, we present a technique, called bypassing, that mitigates the problem. Our technique localizes input memory states only with memory locations that the procedure directly accesses. Those parts not involved in analysis of the procedure are bypassed to transitively called procedures. In experiments with an industrial-strength global C static analyzer, the technique reduces the average analysis time by 42{\%}. In particular, the technique is especially effective for programs that extensively use recursion: it saves analysis time by 77{\%} on average.",
author = "Hakjoo Oh and Kwangkeun Yi",
year = "2011",
month = "12",
day = "26",
doi = "10.1007/978-3-642-25318-8_7",
language = "English",
isbn = "9783642253171",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "50--65",
booktitle = "Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings",

}

TY - GEN

T1 - Access-based localization with bypassing

AU - Oh, Hakjoo

AU - Yi, Kwangkeun

PY - 2011/12/26

Y1 - 2011/12/26

N2 - We present an extension of access-based localization technique to mitigate a substantial inefficiency in handling procedure calls. Recently, access-based localization was proposed as an effective way of tightly localizing abstract memories. However, it has a limitation in handling procedure calls: the localized input memory for a procedure contains not only memory locations accessed by the procedure but also those accessed by transitively called procedures. The weakness is especially exacerbated in the presence of recursive call cycles, which is common in analysis of realistic programs. In this paper, we present a technique, called bypassing, that mitigates the problem. Our technique localizes input memory states only with memory locations that the procedure directly accesses. Those parts not involved in analysis of the procedure are bypassed to transitively called procedures. In experiments with an industrial-strength global C static analyzer, the technique reduces the average analysis time by 42%. In particular, the technique is especially effective for programs that extensively use recursion: it saves analysis time by 77% on average.

AB - We present an extension of access-based localization technique to mitigate a substantial inefficiency in handling procedure calls. Recently, access-based localization was proposed as an effective way of tightly localizing abstract memories. However, it has a limitation in handling procedure calls: the localized input memory for a procedure contains not only memory locations accessed by the procedure but also those accessed by transitively called procedures. The weakness is especially exacerbated in the presence of recursive call cycles, which is common in analysis of realistic programs. In this paper, we present a technique, called bypassing, that mitigates the problem. Our technique localizes input memory states only with memory locations that the procedure directly accesses. Those parts not involved in analysis of the procedure are bypassed to transitively called procedures. In experiments with an industrial-strength global C static analyzer, the technique reduces the average analysis time by 42%. In particular, the technique is especially effective for programs that extensively use recursion: it saves analysis time by 77% on average.

UR - http://www.scopus.com/inward/record.url?scp=84055193118&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84055193118&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-25318-8_7

DO - 10.1007/978-3-642-25318-8_7

M3 - Conference contribution

AN - SCOPUS:84055193118

SN - 9783642253171

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 50

EP - 65

BT - Programming Languages and Systems - 9th Asian Symposium, APLAS 2011, Proceedings

ER -