A study of memory anomalies in openmp applications

Lechen Yu; Joachim Protze; Oscar Hernandez; Vivek Sarkar

doi:10.1007/978-3-030-58144-2_21

A study of memory anomalies in openmp applications

Lechen Yu, Joachim Protze, Oscar Hernandez, Vivek Sarkar

Programming Systems

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Incorrect usage of OpenMP constructs may cause different kinds of defects in OpenMP applications. Most of the existing work focuses on concurrency bugs such as data races and deadlocks, since concurrency bugs are difficult to detect and debug. In this paper, we discuss an under-examined defect in OpenMP applications: memory anomalies. These occur when the application issues illegal memory accesses that may result in a non-deterministic result or even a program crash. Based on the latest OpenMP 5.0 specification, we analyze some OpenMP usage errors that may lead to memory anomalies. Then we illustrate three kinds of memory anomalies: use of uninitialized memory (UUM), use of stale data (USD), and use after free (UAF). While all three anomalies can occur in sequential programs, their manifestations in parallel OpenMP programs can be different, and debugging such anomalies in the context of parallel programs also imposes an additional complexity relative to sequential programs. To measure the effectiveness of memory anomaly detectors on OpenMP applications, we have evaluated three state-of-the-art tools with a group of micro-benchmarks. These micro-benchmarks are either selected from the DRACC benchmark suite or constructed from our own experience. The evaluation result shows that none of these tools can currently handle all three kinds of memory anomalies.

Original language	English
Title of host publication	OpenMP
Subtitle of host publication	Portable Multi-Level Parallelism on Modern Systems - 16th International Workshop on OpenMP, IWOMP 2020, Proceedings
Editors	Kent Milfeld, Lars Koesterke, Bronis R. de Supinski, Jannis Klinkenberg
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	328-342
Number of pages	15
ISBN (Print)	9783030581435
DOIs	https://doi.org/10.1007/978-3-030-58144-2_21
State	Published - 2020
Event	16th International Workshop on OpenMP, IWOMP 2020 - Austin, United States Duration: Sep 22 2020 → Sep 24 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12295 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th International Workshop on OpenMP, IWOMP 2020
Country/Territory	United States
City	Austin
Period	09/22/20 → 09/24/20

Funding

Acknowledgment. This research was supported in part by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of the U.S. Department of Energy Office of Science and the National Nuclear Security Administration, in particular its subproject on Scaling OpenMP with LLVM for Exascale performance and portability (SOLLVE). This work has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement 824080.

Keywords

Memory anomalies
OpenMP
Tools

Access to Document

10.1007/978-3-030-58144-2_21

Cite this

Yu, L., Protze, J., Hernandez, O., & Sarkar, V. (2020). A study of memory anomalies in openmp applications. In K. Milfeld, L. Koesterke, B. R. de Supinski, & J. Klinkenberg (Eds.), OpenMP: Portable Multi-Level Parallelism on Modern Systems - 16th International Workshop on OpenMP, IWOMP 2020, Proceedings (pp. 328-342). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12295 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-58144-2_21

Yu, Lechen ; Protze, Joachim ; Hernandez, Oscar et al. / A study of memory anomalies in openmp applications. OpenMP: Portable Multi-Level Parallelism on Modern Systems - 16th International Workshop on OpenMP, IWOMP 2020, Proceedings. editor / Kent Milfeld ; Lars Koesterke ; Bronis R. de Supinski ; Jannis Klinkenberg. Springer Science and Business Media Deutschland GmbH, 2020. pp. 328-342 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f001ced634a349d0a5276868552afb58,

title = "A study of memory anomalies in openmp applications",

abstract = "Incorrect usage of OpenMP constructs may cause different kinds of defects in OpenMP applications. Most of the existing work focuses on concurrency bugs such as data races and deadlocks, since concurrency bugs are difficult to detect and debug. In this paper, we discuss an under-examined defect in OpenMP applications: memory anomalies. These occur when the application issues illegal memory accesses that may result in a non-deterministic result or even a program crash. Based on the latest OpenMP 5.0 specification, we analyze some OpenMP usage errors that may lead to memory anomalies. Then we illustrate three kinds of memory anomalies: use of uninitialized memory (UUM), use of stale data (USD), and use after free (UAF). While all three anomalies can occur in sequential programs, their manifestations in parallel OpenMP programs can be different, and debugging such anomalies in the context of parallel programs also imposes an additional complexity relative to sequential programs. To measure the effectiveness of memory anomaly detectors on OpenMP applications, we have evaluated three state-of-the-art tools with a group of micro-benchmarks. These micro-benchmarks are either selected from the DRACC benchmark suite or constructed from our own experience. The evaluation result shows that none of these tools can currently handle all three kinds of memory anomalies.",

keywords = "Memory anomalies, OpenMP, Tools",

author = "Lechen Yu and Joachim Protze and Oscar Hernandez and Vivek Sarkar",

note = "Publisher Copyright: {\textcopyright} 2020, The author(s).; 16th International Workshop on OpenMP, IWOMP 2020 ; Conference date: 22-09-2020 Through 24-09-2020",

year = "2020",

doi = "10.1007/978-3-030-58144-2_21",

language = "English",

isbn = "9783030581435",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "328--342",

editor = "Kent Milfeld and Lars Koesterke and {de Supinski}, {Bronis R.} and Jannis Klinkenberg",

booktitle = "OpenMP",

}

Yu, L, Protze, J, Hernandez, O & Sarkar, V 2020, A study of memory anomalies in openmp applications. in K Milfeld, L Koesterke, BR de Supinski & J Klinkenberg (eds), OpenMP: Portable Multi-Level Parallelism on Modern Systems - 16th International Workshop on OpenMP, IWOMP 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12295 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 328-342, 16th International Workshop on OpenMP, IWOMP 2020, Austin, United States, 09/22/20. https://doi.org/10.1007/978-3-030-58144-2_21

A study of memory anomalies in openmp applications. / Yu, Lechen; Protze, Joachim; Hernandez, Oscar et al.
OpenMP: Portable Multi-Level Parallelism on Modern Systems - 16th International Workshop on OpenMP, IWOMP 2020, Proceedings. ed. / Kent Milfeld; Lars Koesterke; Bronis R. de Supinski; Jannis Klinkenberg. Springer Science and Business Media Deutschland GmbH, 2020. p. 328-342 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12295 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A study of memory anomalies in openmp applications

AU - Yu, Lechen

AU - Protze, Joachim

AU - Hernandez, Oscar

AU - Sarkar, Vivek

PY - 2020

Y1 - 2020

N2 - Incorrect usage of OpenMP constructs may cause different kinds of defects in OpenMP applications. Most of the existing work focuses on concurrency bugs such as data races and deadlocks, since concurrency bugs are difficult to detect and debug. In this paper, we discuss an under-examined defect in OpenMP applications: memory anomalies. These occur when the application issues illegal memory accesses that may result in a non-deterministic result or even a program crash. Based on the latest OpenMP 5.0 specification, we analyze some OpenMP usage errors that may lead to memory anomalies. Then we illustrate three kinds of memory anomalies: use of uninitialized memory (UUM), use of stale data (USD), and use after free (UAF). While all three anomalies can occur in sequential programs, their manifestations in parallel OpenMP programs can be different, and debugging such anomalies in the context of parallel programs also imposes an additional complexity relative to sequential programs. To measure the effectiveness of memory anomaly detectors on OpenMP applications, we have evaluated three state-of-the-art tools with a group of micro-benchmarks. These micro-benchmarks are either selected from the DRACC benchmark suite or constructed from our own experience. The evaluation result shows that none of these tools can currently handle all three kinds of memory anomalies.

AB - Incorrect usage of OpenMP constructs may cause different kinds of defects in OpenMP applications. Most of the existing work focuses on concurrency bugs such as data races and deadlocks, since concurrency bugs are difficult to detect and debug. In this paper, we discuss an under-examined defect in OpenMP applications: memory anomalies. These occur when the application issues illegal memory accesses that may result in a non-deterministic result or even a program crash. Based on the latest OpenMP 5.0 specification, we analyze some OpenMP usage errors that may lead to memory anomalies. Then we illustrate three kinds of memory anomalies: use of uninitialized memory (UUM), use of stale data (USD), and use after free (UAF). While all three anomalies can occur in sequential programs, their manifestations in parallel OpenMP programs can be different, and debugging such anomalies in the context of parallel programs also imposes an additional complexity relative to sequential programs. To measure the effectiveness of memory anomaly detectors on OpenMP applications, we have evaluated three state-of-the-art tools with a group of micro-benchmarks. These micro-benchmarks are either selected from the DRACC benchmark suite or constructed from our own experience. The evaluation result shows that none of these tools can currently handle all three kinds of memory anomalies.

KW - Memory anomalies

KW - OpenMP

KW - Tools

UR - http://www.scopus.com/inward/record.url?scp=85091272441&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-58144-2_21

DO - 10.1007/978-3-030-58144-2_21

M3 - Conference contribution

AN - SCOPUS:85091272441

SN - 9783030581435

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

SP - 328

EP - 342

BT - OpenMP

A2 - Milfeld, Kent

A2 - Koesterke, Lars

A2 - de Supinski, Bronis R.

A2 - Klinkenberg, Jannis

PB - Springer Science and Business Media Deutschland GmbH

T2 - 16th International Workshop on OpenMP, IWOMP 2020

Y2 - 22 September 2020 through 24 September 2020

ER -

Yu L, Protze J, Hernandez O, Sarkar V. A study of memory anomalies in openmp applications. In Milfeld K, Koesterke L, de Supinski BR, Klinkenberg J, editors, OpenMP: Portable Multi-Level Parallelism on Modern Systems - 16th International Workshop on OpenMP, IWOMP 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 328-342. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-58144-2_21

A study of memory anomalies in openmp applications

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this