BLAS-3 Optimized by OmpSs Regions (LASs Library)

Pedro Valero-Lara; Sandra Catalán; Xavier Martorell; Jesús Labarta

doi:10.1109/EMPDP.2019.8671545

BLAS-3 Optimized by OmpSs Regions (LASs Library)

Pedro Valero-Lara, Sandra Catalán, Xavier Martorell, Jesús Labarta

Programming Systems

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

10 Scopus citations

Abstract

In this paper we propose a set of optimizations for the BLAS-3 routines of LASs library (Linear Algebra routines on OmpSs) and perform a detailed analysis of the impact of the proposed changes in terms of performance and execution time. OmpSs allows to use regions in the dependences of the tasks. This helps not only in the programming of the algorithmic optimizations, but also in the reduction of the execution time achieved by such optimizations. Different strategies are implemented in order to reduce the amount of tasks created (when there is enough parallelism) during the execution of BLAS-3 operations in the original LASs. Also a better IPC is obtained thanks to a better memory hierarchy exploitation. More specifically, we increase the performance, in particular on big matrices, about 12% for TRSM, and 17% for GEMM with respect to the original version of LASs, even using less cores in the case of GEMM/SYMM. Moreover, when LASs is compared to the OpenMP reference dense linear algebra library PLASMA, performance is increased up to 12.5% for GEMM/SYMM, while for TRSM/TRMM this value raises to 15%.

Original language	English
Title of host publication	Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	25-32
Number of pages	8
ISBN (Electronic)	9781728116440
DOIs	https://doi.org/10.1109/EMPDP.2019.8671545
State	Published - Mar 19 2019
Event	27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019 - Pavia, Italy Duration: Feb 13 2019 → Feb 15 2019

Publication series

Name	Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019

Conference

Conference	27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019
Country/Territory	Italy
City	Pavia
Period	02/13/19 → 02/15/19

Funding

ACKNOWLEDGMENT This project has received funding from the Spanish Ministry of Economy and Competitiveness under the project Computación de Altas Prestaciones VII (TIN2015-65316-P), the Departament d’Innovació, Universitats i Empresa de la Generalitat de Catalunya, under project MPEXPAR: Models de Programació i Entorns d’Execució Parallels (2014-SGR-1051), and the Juan de la Cierva Grant Agreement No IJCI-2017-33511. We also acknowledge the funding provided by Fujitsu under the BSC-Fujitsu joint project: Math Libraries Migration and Optimization.

Keywords

BLAS-3
OmpSs
regions
tasking

Access to Document

10.1109/EMPDP.2019.8671545

Cite this

Valero-Lara, P., Catalán, S., Martorell, X., & Labarta, J. (2019). BLAS-3 Optimized by OmpSs Regions (LASs Library). In Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019 (pp. 25-32). Article 8671545 (Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EMPDP.2019.8671545

Valero-Lara, Pedro ; Catalán, Sandra ; Martorell, Xavier et al. / BLAS-3 Optimized by OmpSs Regions (LASs Library). Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 25-32 (Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019).

@inproceedings{20821e0377604797b17e800928fd8d6c,

title = "BLAS-3 Optimized by OmpSs Regions (LASs Library)",

abstract = "In this paper we propose a set of optimizations for the BLAS-3 routines of LASs library (Linear Algebra routines on OmpSs) and perform a detailed analysis of the impact of the proposed changes in terms of performance and execution time. OmpSs allows to use regions in the dependences of the tasks. This helps not only in the programming of the algorithmic optimizations, but also in the reduction of the execution time achieved by such optimizations. Different strategies are implemented in order to reduce the amount of tasks created (when there is enough parallelism) during the execution of BLAS-3 operations in the original LASs. Also a better IPC is obtained thanks to a better memory hierarchy exploitation. More specifically, we increase the performance, in particular on big matrices, about 12% for TRSM, and 17% for GEMM with respect to the original version of LASs, even using less cores in the case of GEMM/SYMM. Moreover, when LASs is compared to the OpenMP reference dense linear algebra library PLASMA, performance is increased up to 12.5% for GEMM/SYMM, while for TRSM/TRMM this value raises to 15%.",

keywords = "BLAS-3, OmpSs, regions, tasking",

author = "Pedro Valero-Lara and Sandra Catal{\'a}n and Xavier Martorell and Jes{\'u}s Labarta",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019 ; Conference date: 13-02-2019 Through 15-02-2019",

year = "2019",

month = mar,

day = "19",

doi = "10.1109/EMPDP.2019.8671545",

language = "English",

series = "Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "25--32",

booktitle = "Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019",

}

Valero-Lara, P, Catalán, S, Martorell, X & Labarta, J 2019, BLAS-3 Optimized by OmpSs Regions (LASs Library). in Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019., 8671545, Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019, Institute of Electrical and Electronics Engineers Inc., pp. 25-32, 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019, Pavia, Italy, 02/13/19. https://doi.org/10.1109/EMPDP.2019.8671545

BLAS-3 Optimized by OmpSs Regions (LASs Library). / Valero-Lara, Pedro; Catalán, Sandra; Martorell, Xavier et al.
Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 25-32 8671545 (Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019).

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

TY - GEN

T1 - BLAS-3 Optimized by OmpSs Regions (LASs Library)

AU - Valero-Lara, Pedro

AU - Catalán, Sandra

AU - Martorell, Xavier

AU - Labarta, Jesús

PY - 2019/3/19

Y1 - 2019/3/19

N2 - In this paper we propose a set of optimizations for the BLAS-3 routines of LASs library (Linear Algebra routines on OmpSs) and perform a detailed analysis of the impact of the proposed changes in terms of performance and execution time. OmpSs allows to use regions in the dependences of the tasks. This helps not only in the programming of the algorithmic optimizations, but also in the reduction of the execution time achieved by such optimizations. Different strategies are implemented in order to reduce the amount of tasks created (when there is enough parallelism) during the execution of BLAS-3 operations in the original LASs. Also a better IPC is obtained thanks to a better memory hierarchy exploitation. More specifically, we increase the performance, in particular on big matrices, about 12% for TRSM, and 17% for GEMM with respect to the original version of LASs, even using less cores in the case of GEMM/SYMM. Moreover, when LASs is compared to the OpenMP reference dense linear algebra library PLASMA, performance is increased up to 12.5% for GEMM/SYMM, while for TRSM/TRMM this value raises to 15%.

AB - In this paper we propose a set of optimizations for the BLAS-3 routines of LASs library (Linear Algebra routines on OmpSs) and perform a detailed analysis of the impact of the proposed changes in terms of performance and execution time. OmpSs allows to use regions in the dependences of the tasks. This helps not only in the programming of the algorithmic optimizations, but also in the reduction of the execution time achieved by such optimizations. Different strategies are implemented in order to reduce the amount of tasks created (when there is enough parallelism) during the execution of BLAS-3 operations in the original LASs. Also a better IPC is obtained thanks to a better memory hierarchy exploitation. More specifically, we increase the performance, in particular on big matrices, about 12% for TRSM, and 17% for GEMM with respect to the original version of LASs, even using less cores in the case of GEMM/SYMM. Moreover, when LASs is compared to the OpenMP reference dense linear algebra library PLASMA, performance is increased up to 12.5% for GEMM/SYMM, while for TRSM/TRMM this value raises to 15%.

KW - BLAS-3

KW - OmpSs

KW - regions

KW - tasking

UR - http://www.scopus.com/inward/record.url?scp=85063866967&partnerID=8YFLogxK

U2 - 10.1109/EMPDP.2019.8671545

DO - 10.1109/EMPDP.2019.8671545

M3 - Conference contribution

AN - SCOPUS:85063866967

T3 - Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019

SP - 25

EP - 32

BT - Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019

Y2 - 13 February 2019 through 15 February 2019

ER -

Valero-Lara P, Catalán S, Martorell X, Labarta J. BLAS-3 Optimized by OmpSs Regions (LASs Library). In Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 25-32. 8671545. (Proceedings - 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, PDP 2019). doi: 10.1109/EMPDP.2019.8671545

BLAS-3 Optimized by OmpSs Regions (LASs Library)

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this