TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers

Simon McIntosh-Smith; Matt Martineau; Tom Deakin; Grzegorz Pawelczak; Wayne Gaudin; Paul Garrett; Wei Liu; Richard Smedley-Stevenson; David Beckingsale

doi:10.1109/CLUSTER.2017.105

TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers

Simon McIntosh-Smith, Matt Martineau, Tom Deakin, Grzegorz Pawelczak, Wayne Gaudin, Paul Garrett, Wei Liu, Richard Smedley-Stevenson, David Beckingsale

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

11 Citations (Scopus)

443 Downloads (Pure)

Abstract

Iterative sparse linear solvers are an important class of algorithm in high performance computing, and form a crucial component of many scientific codes. As intra and inter node parallelism continues to increase rapidly, the design of new, scalable solvers which can target next generation architectures becomes increasingly important. In this work we present TeaLeaf, a recent mini-App constructed to explore design space choices for highly scalable solvers. We then use TeaLeaf to compare the standard CG algorithm with a Chebyshev Polynomially Preconditioned Conjugate Gradient (CPPCG) iterative sparse linear solver. CPPCG is a communication-Avoiding algorithm, requiring less global communication than previous approaches. TeaLeaf includes support for many-core processors, such as GPUS and Xeon Phi, and we include strong-scaling results across a range of world-leading Petascale supercomputers, including Titan and Piz Daint.

Original language	English
Title of host publication	2017 IEEE International Conference on Cluster Computing (CLUSTER 2017)
Subtitle of host publication	Proceedings of a meeting held 5-8 September 2017, Honolulu, Hawaii, USA
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	842-849
Number of pages	8
ISBN (Electronic)	9781538623268
ISBN (Print)	9781538623275
DOIs	https://doi.org/10.1109/CLUSTER.2017.105
Publication status	Published - Oct 2017
Event	2017 IEEE International Conference on Cluster Computing, CLUSTER 2017 - Honolulu, United States Duration: 5 Sep 2017 → 8 Sep 2017

Publication series

Name
ISSN (Print)	2168-9253

Conference

Conference	2017 IEEE International Conference on Cluster Computing, CLUSTER 2017
Country/Territory	United States
City	Honolulu
Period	5/09/17 → 8/09/17

Keywords

Accelerators
Iterative parse linear solvers
Mini-Apps
Strong scaling

Access to Document

10.1109/CLUSTER.2017.105

Full-text PDF (accepted author manuscript)
This is the author accepted manuscript (AAM). The final published version (version of record) is available online via IEEE at http://ieeexplore.ieee.org/document/8049027/ . Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 284 KB

Cite this

McIntosh-Smith, S., Martineau, M., Deakin, T., Pawelczak, G., Gaudin, W., Garrett, P., Liu, W., Smedley-Stevenson, R., & Beckingsale, D. (2017). TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers. In 2017 IEEE International Conference on Cluster Computing (CLUSTER 2017): Proceedings of a meeting held 5-8 September 2017, Honolulu, Hawaii, USA (pp. 842-849). [8049027] Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/CLUSTER.2017.105

McIntosh-Smith, Simon ; Martineau, Matt ; Deakin, Tom ; Pawelczak, Grzegorz ; Gaudin, Wayne ; Garrett, Paul ; Liu, Wei ; Smedley-Stevenson, Richard ; Beckingsale, David. / TeaLeaf : A mini-application to enable design-space explorations for iterative sparse linear solvers. 2017 IEEE International Conference on Cluster Computing (CLUSTER 2017): Proceedings of a meeting held 5-8 September 2017, Honolulu, Hawaii, USA. Institute of Electrical and Electronics Engineers (IEEE), 2017. pp. 842-849

@inproceedings{2e9df5995fa94ad09fa6234035a2cc0e,

title = "TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers",

abstract = "Iterative sparse linear solvers are an important class of algorithm in high performance computing, and form a crucial component of many scientific codes. As intra and inter node parallelism continues to increase rapidly, the design of new, scalable solvers which can target next generation architectures becomes increasingly important. In this work we present TeaLeaf, a recent mini-App constructed to explore design space choices for highly scalable solvers. We then use TeaLeaf to compare the standard CG algorithm with a Chebyshev Polynomially Preconditioned Conjugate Gradient (CPPCG) iterative sparse linear solver. CPPCG is a communication-Avoiding algorithm, requiring less global communication than previous approaches. TeaLeaf includes support for many-core processors, such as GPUS and Xeon Phi, and we include strong-scaling results across a range of world-leading Petascale supercomputers, including Titan and Piz Daint.",

keywords = "Accelerators, Iterative parse linear solvers, Mini-Apps, Strong scaling",

author = "Simon McIntosh-Smith and Matt Martineau and Tom Deakin and Grzegorz Pawelczak and Wayne Gaudin and Paul Garrett and Wei Liu and Richard Smedley-Stevenson and David Beckingsale",

year = "2017",

month = oct,

doi = "10.1109/CLUSTER.2017.105",

language = "English",

isbn = "9781538623275",

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

pages = "842--849",

booktitle = "2017 IEEE International Conference on Cluster Computing (CLUSTER 2017)",

address = "United States",

note = "2017 IEEE International Conference on Cluster Computing, CLUSTER 2017 ; Conference date: 05-09-2017 Through 08-09-2017",

}

McIntosh-Smith, S, Martineau, M, Deakin, T, Pawelczak, G, Gaudin, W, Garrett, P, Liu, W, Smedley-Stevenson, R & Beckingsale, D 2017, TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers. in 2017 IEEE International Conference on Cluster Computing (CLUSTER 2017): Proceedings of a meeting held 5-8 September 2017, Honolulu, Hawaii, USA., 8049027, Institute of Electrical and Electronics Engineers (IEEE), pp. 842-849, 2017 IEEE International Conference on Cluster Computing, CLUSTER 2017, Honolulu, Hawaii, United States, 5/09/17. https://doi.org/10.1109/CLUSTER.2017.105

TeaLeaf : A mini-application to enable design-space explorations for iterative sparse linear solvers. / McIntosh-Smith, Simon; Martineau, Matt; Deakin, Tom; Pawelczak, Grzegorz; Gaudin, Wayne; Garrett, Paul; Liu, Wei; Smedley-Stevenson, Richard; Beckingsale, David.

2017 IEEE International Conference on Cluster Computing (CLUSTER 2017): Proceedings of a meeting held 5-8 September 2017, Honolulu, Hawaii, USA. Institute of Electrical and Electronics Engineers (IEEE), 2017. p. 842-849 8049027.

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - TeaLeaf

T2 - 2017 IEEE International Conference on Cluster Computing, CLUSTER 2017

AU - McIntosh-Smith, Simon

AU - Martineau, Matt

AU - Deakin, Tom

AU - Pawelczak, Grzegorz

AU - Gaudin, Wayne

AU - Garrett, Paul

AU - Liu, Wei

AU - Smedley-Stevenson, Richard

AU - Beckingsale, David

PY - 2017/10

Y1 - 2017/10

N2 - Iterative sparse linear solvers are an important class of algorithm in high performance computing, and form a crucial component of many scientific codes. As intra and inter node parallelism continues to increase rapidly, the design of new, scalable solvers which can target next generation architectures becomes increasingly important. In this work we present TeaLeaf, a recent mini-App constructed to explore design space choices for highly scalable solvers. We then use TeaLeaf to compare the standard CG algorithm with a Chebyshev Polynomially Preconditioned Conjugate Gradient (CPPCG) iterative sparse linear solver. CPPCG is a communication-Avoiding algorithm, requiring less global communication than previous approaches. TeaLeaf includes support for many-core processors, such as GPUS and Xeon Phi, and we include strong-scaling results across a range of world-leading Petascale supercomputers, including Titan and Piz Daint.

AB - Iterative sparse linear solvers are an important class of algorithm in high performance computing, and form a crucial component of many scientific codes. As intra and inter node parallelism continues to increase rapidly, the design of new, scalable solvers which can target next generation architectures becomes increasingly important. In this work we present TeaLeaf, a recent mini-App constructed to explore design space choices for highly scalable solvers. We then use TeaLeaf to compare the standard CG algorithm with a Chebyshev Polynomially Preconditioned Conjugate Gradient (CPPCG) iterative sparse linear solver. CPPCG is a communication-Avoiding algorithm, requiring less global communication than previous approaches. TeaLeaf includes support for many-core processors, such as GPUS and Xeon Phi, and we include strong-scaling results across a range of world-leading Petascale supercomputers, including Titan and Piz Daint.

KW - Accelerators

KW - Iterative parse linear solvers

KW - Mini-Apps

KW - Strong scaling

UR - http://www.scopus.com/inward/record.url?scp=85032639782&partnerID=8YFLogxK

U2 - 10.1109/CLUSTER.2017.105

DO - 10.1109/CLUSTER.2017.105

M3 - Conference Contribution (Conference Proceeding)

AN - SCOPUS:85032639782

SN - 9781538623275

SP - 842

EP - 849

BT - 2017 IEEE International Conference on Cluster Computing (CLUSTER 2017)

PB - Institute of Electrical and Electronics Engineers (IEEE)

Y2 - 5 September 2017 through 8 September 2017

ER -

McIntosh-Smith S, Martineau M, Deakin T, Pawelczak G, Gaudin W, Garrett P et al. TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers. In 2017 IEEE International Conference on Cluster Computing (CLUSTER 2017): Proceedings of a meeting held 5-8 September 2017, Honolulu, Hawaii, USA. Institute of Electrical and Electronics Engineers (IEEE). 2017. p. 842-849. 8049027 https://doi.org/10.1109/CLUSTER.2017.105

TeaLeaf: A mini-application to enable design-space explorations for iterative sparse linear solvers

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this