X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs

Sven Rheindt; Andreas Fried; Oliver Lenke; Lars Nolte; Temur Sabirov; Tim Twardzik; Thomas Wild; Andreas Herkersdorf

doi:10.1007/978-3-030-52794-5_9

X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs

Sven Rheindt, Andreas Fried, Oliver Lenke, Lars Nolte, Temur Sabirov, Tim Twardzik, Thomas Wild, Andreas Herkersdorf

Chair of Integrated Systems

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

1 Scopus citations

Abstract

Near-memory acceleration strives to tackle the data-to-task locality issue in MPSoCs in order to obtain higher performance and lower power consumption. However, it is not easy to determine whether the advantages arise from the near-memory integration or the hardware acceleration (versus software execution). We propose X-CEL, a method to accurately estimate the potential of near-memory acceleration using an easy-to-integrate near-memory core. We showcase X-CEL’s benefits with three variants of graph copy mechanisms in a tile-based MPSoC. Evaluations reveal that the estimated speedup is in good accordance with the actual speedup achieved by the near-memory accelerator.

Original language	English
Title of host publication	Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings
Editors	André Brinkmann, Wolfgang Karl, Stefan Lankes, Sven Tomforde, Thilo Pionteck, Carsten Trinitis
Publisher	Springer
Pages	109-123
Number of pages	15
ISBN (Print)	9783030527938
DOIs	https://doi.org/10.1007/978-3-030-52794-5_9
State	Published - 2020
Event	33rd International Conference on Architecture of Computing Systems, ARCS 2020 - Aachen, Germany Duration: 25 May 2020 → 28 May 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12155 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	33rd International Conference on Architecture of Computing Systems, ARCS 2020
Country/Territory	Germany
City	Aachen
Period	25/05/20 → 28/05/20

Keywords

Data-to-task locality
Design space exploration
Graph copy
Near-memory acceleration
Tile-based MPSoC

Access to Document

10.1007/978-3-030-52794-5_9

Cite this

Rheindt, S., Fried, A., Lenke, O., Nolte, L., Sabirov, T., Twardzik, T., Wild, T., & Herkersdorf, A. (2020). X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs. In A. Brinkmann, W. Karl, S. Lankes, S. Tomforde, T. Pionteck, & C. Trinitis (Eds.), Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings (pp. 109-123). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12155 LNCS). Springer. https://doi.org/10.1007/978-3-030-52794-5_9

Rheindt, Sven ; Fried, Andreas ; Lenke, Oliver et al. / X-cel : A method to estimate near-memory acceleration potential in tile-based mpsocs. Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings. editor / André Brinkmann ; Wolfgang Karl ; Stefan Lankes ; Sven Tomforde ; Thilo Pionteck ; Carsten Trinitis. Springer, 2020. pp. 109-123 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{6185af6ef80c4e3fa6a3dcfbc0020b94,

title = "X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs",

abstract = "Near-memory acceleration strives to tackle the data-to-task locality issue in MPSoCs in order to obtain higher performance and lower power consumption. However, it is not easy to determine whether the advantages arise from the near-memory integration or the hardware acceleration (versus software execution). We propose X-CEL, a method to accurately estimate the potential of near-memory acceleration using an easy-to-integrate near-memory core. We showcase X-CEL{\textquoteright}s benefits with three variants of graph copy mechanisms in a tile-based MPSoC. Evaluations reveal that the estimated speedup is in good accordance with the actual speedup achieved by the near-memory accelerator.",

keywords = "Data-to-task locality, Design space exploration, Graph copy, Near-memory acceleration, Tile-based MPSoC",

author = "Sven Rheindt and Andreas Fried and Oliver Lenke and Lars Nolte and Temur Sabirov and Tim Twardzik and Thomas Wild and Andreas Herkersdorf",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2020.; 33rd International Conference on Architecture of Computing Systems, ARCS 2020 ; Conference date: 25-05-2020 Through 28-05-2020",

year = "2020",

doi = "10.1007/978-3-030-52794-5_9",

language = "English",

isbn = "9783030527938",

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

publisher = "Springer",

pages = "109--123",

editor = "Andr{\'e} Brinkmann and Wolfgang Karl and Stefan Lankes and Sven Tomforde and Thilo Pionteck and Carsten Trinitis",

booktitle = "Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings",

}

Rheindt, S, Fried, A, Lenke, O, Nolte, L, Sabirov, T, Twardzik, T, Wild, T & Herkersdorf, A 2020, X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs. in A Brinkmann, W Karl, S Lankes, S Tomforde, T Pionteck & C Trinitis (eds), Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12155 LNCS, Springer, pp. 109-123, 33rd International Conference on Architecture of Computing Systems, ARCS 2020, Aachen, Germany, 25/05/20. https://doi.org/10.1007/978-3-030-52794-5_9

X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs. / Rheindt, Sven; Fried, Andreas; Lenke, Oliver et al.
Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings. ed. / André Brinkmann; Wolfgang Karl; Stefan Lankes; Sven Tomforde; Thilo Pionteck; Carsten Trinitis. Springer, 2020. p. 109-123 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12155 LNCS).

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

TY - GEN

T1 - X-cel

T2 - 33rd International Conference on Architecture of Computing Systems, ARCS 2020

AU - Rheindt, Sven

AU - Fried, Andreas

AU - Lenke, Oliver

AU - Nolte, Lars

AU - Sabirov, Temur

AU - Twardzik, Tim

AU - Wild, Thomas

AU - Herkersdorf, Andreas

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2020.

PY - 2020

Y1 - 2020

N2 - Near-memory acceleration strives to tackle the data-to-task locality issue in MPSoCs in order to obtain higher performance and lower power consumption. However, it is not easy to determine whether the advantages arise from the near-memory integration or the hardware acceleration (versus software execution). We propose X-CEL, a method to accurately estimate the potential of near-memory acceleration using an easy-to-integrate near-memory core. We showcase X-CEL’s benefits with three variants of graph copy mechanisms in a tile-based MPSoC. Evaluations reveal that the estimated speedup is in good accordance with the actual speedup achieved by the near-memory accelerator.

AB - Near-memory acceleration strives to tackle the data-to-task locality issue in MPSoCs in order to obtain higher performance and lower power consumption. However, it is not easy to determine whether the advantages arise from the near-memory integration or the hardware acceleration (versus software execution). We propose X-CEL, a method to accurately estimate the potential of near-memory acceleration using an easy-to-integrate near-memory core. We showcase X-CEL’s benefits with three variants of graph copy mechanisms in a tile-based MPSoC. Evaluations reveal that the estimated speedup is in good accordance with the actual speedup achieved by the near-memory accelerator.

KW - Data-to-task locality

KW - Design space exploration

KW - Graph copy

KW - Near-memory acceleration

KW - Tile-based MPSoC

UR - http://www.scopus.com/inward/record.url?scp=85088751310&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-52794-5_9

DO - 10.1007/978-3-030-52794-5_9

M3 - Conference contribution

AN - SCOPUS:85088751310

SN - 9783030527938

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

SP - 109

EP - 123

BT - Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings

A2 - Brinkmann, André

A2 - Karl, Wolfgang

A2 - Lankes, Stefan

A2 - Tomforde, Sven

A2 - Pionteck, Thilo

A2 - Trinitis, Carsten

PB - Springer

Y2 - 25 May 2020 through 28 May 2020

ER -

Rheindt S, Fried A, Lenke O, Nolte L, Sabirov T, Twardzik T et al. X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs. In Brinkmann A, Karl W, Lankes S, Tomforde S, Pionteck T, Trinitis C, editors, Architecture of Computing Systems – ARCS 2020 - 33rd International Conference, Proceedings. Springer. 2020. p. 109-123. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-52794-5_9

X-cel: A method to estimate near-memory acceleration potential in tile-based mpsocs

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this