A Hybrid Approach to Parallel Pattern Discovery in C++

C Brown; V. Janjic; A. Barwell; J. Thomson; Roberto Castaneda Lozano; Murray Cole; Bjoern Franke; J.D. Garcia-Sanchez; D. Del Rio Astorga; K. MacKenzie

doi:10.1109/PDP50117.2020.00035

A Hybrid Approach to Parallel Pattern Discovery in C++

C Brown, V. Janjic, A. Barwell, J. Thomson, Roberto Castaneda Lozano, Murray Cole, Bjoern Franke, J.D. Garcia-Sanchez, D. Del Rio Astorga, K. MacKenzie

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

Abstract

Parallel pattern libraries offer a strong combination of abstraction and performance. However, discovering places in sequential code where parallel patterns should be introduced is still highly non-trivial, often requiring expert manual analysis and profiling. We present a hybrid discovery technique to detect instances of parallel patterns in sequential code. This employs both static and dynamic trace-based analysis, together with hotspot detection. We evaluate our pattern discovery mechanism on a number of representative benchmarks. We evaluate the performance of the resulting parallelised benchmarks on a 24-core parallel machine.

Original language	English
Title of host publication	2020 28th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	187-191
Number of pages	5
ISBN (Electronic)	978-1-7281-6582-0
ISBN (Print)	978-1-7281-6583-7
DOIs	https://doi.org/10.1109/PDP50117.2020.00035
Publication status	Published - 14 May 2020
Event	28th Euromicro International Conference on Parallel, Distributed and Network-base Processing - Västerås, Sweden Duration: 11 Mar 2020 → 13 Mar 2020 http://www.pdp2020.com/

Publication series

Name
Publisher	IEEE
ISSN (Print)	1066-6192
ISSN (Electronic)	2377-5750

Conference

Conference	28th Euromicro International Conference on Parallel, Distributed and Network-base Processing
Abbreviated title	PDP 2020
Country/Territory	Sweden
City	Västerås
Period	11/03/20 → 13/03/20
Internet address	http://www.pdp2020.com/

Access to Document

10.1109/PDP50117.2020.00035Licence: All Rights Reserved

A Hybrid Approach_BROWN_DOA01122019_AFVAccepted author manuscript, 186 KB

https://ieeexplore.ieee.org/document/9092377Licence: All Rights Reserved

Cite this

Brown, C., Janjic, V., Barwell, A., Thomson, J., Castaneda Lozano, R., Cole, M., Franke, B., Garcia-Sanchez, J. D., Del Rio Astorga, D., & MacKenzie, K. (2020). A Hybrid Approach to Parallel Pattern Discovery in C++. In 2020 28th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP) (pp. 187-191). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/PDP50117.2020.00035

@inproceedings{7fafd27cbfda4c16a0ca364386e1143e,

title = "A Hybrid Approach to Parallel Pattern Discovery in C++",

abstract = "Parallel pattern libraries offer a strong combination of abstraction and performance. However, discovering places in sequential code where parallel patterns should be introduced is still highly non-trivial, often requiring expert manual analysis and profiling. We present a hybrid discovery technique to detect instances of parallel patterns in sequential code. This employs both static and dynamic trace-based analysis, together with hotspot detection. We evaluate our pattern discovery mechanism on a number of representative benchmarks. We evaluate the performance of the resulting parallelised benchmarks on a 24-core parallel machine.",

author = "C Brown and V. Janjic and A. Barwell and J. Thomson and {Castaneda Lozano}, Roberto and Murray Cole and Bjoern Franke and J.D. Garcia-Sanchez and {Del Rio Astorga}, D. and K. MacKenzie",

year = "2020",

month = may,

day = "14",

doi = "10.1109/PDP50117.2020.00035",

language = "English",

isbn = "978-1-7281-6583-7",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

pages = "187--191",

booktitle = "2020 28th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP)",

address = "United States",

note = "28th Euromicro International Conference on Parallel, Distributed and Network-base Processing, PDP 2020 ; Conference date: 11-03-2020 Through 13-03-2020",

url = "http://www.pdp2020.com/",

}

Brown, C, Janjic, V, Barwell, A, Thomson, J, Castaneda Lozano, R, Cole, M , Franke, B, Garcia-Sanchez, JD, Del Rio Astorga, D & MacKenzie, K 2020, A Hybrid Approach to Parallel Pattern Discovery in C++. in 2020 28th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP). Institute of Electrical and Electronics Engineers (IEEE), pp. 187-191, 28th Euromicro International Conference on Parallel, Distributed and Network-base Processing, Västerås, Sweden, 11/03/20. https://doi.org/10.1109/PDP50117.2020.00035

TY - GEN

T1 - A Hybrid Approach to Parallel Pattern Discovery in C++

AU - Brown, C

AU - Janjic, V.

AU - Barwell, A.

AU - Thomson, J.

AU - Castaneda Lozano, Roberto

AU - Cole, Murray

AU - Franke, Bjoern

AU - Garcia-Sanchez, J.D.

AU - Del Rio Astorga, D.

AU - MacKenzie, K.

PY - 2020/5/14

Y1 - 2020/5/14

N2 - Parallel pattern libraries offer a strong combination of abstraction and performance. However, discovering places in sequential code where parallel patterns should be introduced is still highly non-trivial, often requiring expert manual analysis and profiling. We present a hybrid discovery technique to detect instances of parallel patterns in sequential code. This employs both static and dynamic trace-based analysis, together with hotspot detection. We evaluate our pattern discovery mechanism on a number of representative benchmarks. We evaluate the performance of the resulting parallelised benchmarks on a 24-core parallel machine.

AB - Parallel pattern libraries offer a strong combination of abstraction and performance. However, discovering places in sequential code where parallel patterns should be introduced is still highly non-trivial, often requiring expert manual analysis and profiling. We present a hybrid discovery technique to detect instances of parallel patterns in sequential code. This employs both static and dynamic trace-based analysis, together with hotspot detection. We evaluate our pattern discovery mechanism on a number of representative benchmarks. We evaluate the performance of the resulting parallelised benchmarks on a 24-core parallel machine.

U2 - 10.1109/PDP50117.2020.00035

DO - 10.1109/PDP50117.2020.00035

M3 - Conference contribution

SN - 978-1-7281-6583-7

SP - 187

EP - 191

BT - 2020 28th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP)

PB - Institute of Electrical and Electronics Engineers (IEEE)

T2 - 28th Euromicro International Conference on Parallel, Distributed and Network-base Processing

Y2 - 11 March 2020 through 13 March 2020

ER -