Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures

Claudio Machado Diniz; Muhammad Shafique; Sergio Bampi; Jorg Henkel

doi:10.1109/FPL.2014.6927392

Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures

Claudio Machado Diniz, Muhammad Shafique, Sergio Bampi, Jorg Henkel

Electrical Engineering

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

Abstract

Run-time mixed-grained reconfigurable architectures emerged as an efficient solution to deal with the heterogeneous and at-design-time unpredictable nature of advanced applications. Due to interconnection limitations, the reconfigurable elements are grouped into tiles communicating through an on-chip network. State-of-the-art run-time accelerator binding schemes, i.e., mapping the accelerators to elements in the physical reconfigurable array, do not deal with such tile-based architectures. We propose a new scheme for run-time accelerator binding into our tile-based mixed-grained reconfigurable architecture. By means of an advanced video encoding application, we illustrate that our scheme reduces the inter-tile communication overhead by up to 44% (avg. 23%).

Original language	English (US)
Title of host publication	Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9783000446450
DOIs	https://doi.org/10.1109/FPL.2014.6927392
State	Published - Oct 16 2014
Event	24th International Conference on Field Programmable Logic and Applications, FPL 2014 - Munich, Germany Duration: Sep 1 2014 → Sep 5 2014

Publication series

Name	Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014

Conference

Conference	24th International Conference on Field Programmable Logic and Applications, FPL 2014
Country/Territory	Germany
City	Munich
Period	9/1/14 → 9/5/14

Keywords

Binding
Reconfigurable architecture
Tiles

ASJC Scopus subject areas

Computer Science Applications
Hardware and Architecture

Access to Document

10.1109/FPL.2014.6927392

Cite this

Diniz, C. M., Shafique, M., Bampi, S., & Henkel, J. (2014). Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures. In Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014 Article 6927392 (Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/FPL.2014.6927392

Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures. / Diniz, Claudio Machado; Shafique, Muhammad; Bampi, Sergio et al.
Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014. Institute of Electrical and Electronics Engineers Inc., 2014. 6927392 (Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014).

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

Diniz, CM, Shafique, M, Bampi, S & Henkel, J 2014, Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures. in Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014., 6927392, Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014, Institute of Electrical and Electronics Engineers Inc., 24th International Conference on Field Programmable Logic and Applications, FPL 2014, Munich, Germany, 9/1/14. https://doi.org/10.1109/FPL.2014.6927392

Diniz CM, Shafique M, Bampi S, Henkel J. Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures. In Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014. Institute of Electrical and Electronics Engineers Inc. 2014. 6927392. (Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014). doi: 10.1109/FPL.2014.6927392

Diniz, Claudio Machado ; Shafique, Muhammad ; Bampi, Sergio et al. / Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures. Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014. Institute of Electrical and Electronics Engineers Inc., 2014. (Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014).

@inproceedings{d4b2841962b44e0ba47f669392617c09,

title = "Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures",

abstract = "Run-time mixed-grained reconfigurable architectures emerged as an efficient solution to deal with the heterogeneous and at-design-time unpredictable nature of advanced applications. Due to interconnection limitations, the reconfigurable elements are grouped into tiles communicating through an on-chip network. State-of-the-art run-time accelerator binding schemes, i.e., mapping the accelerators to elements in the physical reconfigurable array, do not deal with such tile-based architectures. We propose a new scheme for run-time accelerator binding into our tile-based mixed-grained reconfigurable architecture. By means of an advanced video encoding application, we illustrate that our scheme reduces the inter-tile communication overhead by up to 44% (avg. 23%).",

keywords = "Binding, Reconfigurable architecture, Tiles",

author = "Diniz, {Claudio Machado} and Muhammad Shafique and Sergio Bampi and Jorg Henkel",

note = "Publisher Copyright: {\textcopyright} 2014 Technical University of Munich (TUM). Copyright: Copyright 2014 Elsevier B.V., All rights reserved.; 24th International Conference on Field Programmable Logic and Applications, FPL 2014 ; Conference date: 01-09-2014 Through 05-09-2014",

year = "2014",

month = oct,

day = "16",

doi = "10.1109/FPL.2014.6927392",

language = "English (US)",

series = "Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014",

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

booktitle = "Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014",

}

TY - GEN

T1 - Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures

AU - Diniz, Claudio Machado

AU - Shafique, Muhammad

AU - Bampi, Sergio

AU - Henkel, Jorg

PY - 2014/10/16

Y1 - 2014/10/16

N2 - Run-time mixed-grained reconfigurable architectures emerged as an efficient solution to deal with the heterogeneous and at-design-time unpredictable nature of advanced applications. Due to interconnection limitations, the reconfigurable elements are grouped into tiles communicating through an on-chip network. State-of-the-art run-time accelerator binding schemes, i.e., mapping the accelerators to elements in the physical reconfigurable array, do not deal with such tile-based architectures. We propose a new scheme for run-time accelerator binding into our tile-based mixed-grained reconfigurable architecture. By means of an advanced video encoding application, we illustrate that our scheme reduces the inter-tile communication overhead by up to 44% (avg. 23%).

AB - Run-time mixed-grained reconfigurable architectures emerged as an efficient solution to deal with the heterogeneous and at-design-time unpredictable nature of advanced applications. Due to interconnection limitations, the reconfigurable elements are grouped into tiles communicating through an on-chip network. State-of-the-art run-time accelerator binding schemes, i.e., mapping the accelerators to elements in the physical reconfigurable array, do not deal with such tile-based architectures. We propose a new scheme for run-time accelerator binding into our tile-based mixed-grained reconfigurable architecture. By means of an advanced video encoding application, we illustrate that our scheme reduces the inter-tile communication overhead by up to 44% (avg. 23%).

KW - Binding

KW - Reconfigurable architecture

KW - Tiles

UR - http://www.scopus.com/inward/record.url?scp=84911165172&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84911165172&partnerID=8YFLogxK

U2 - 10.1109/FPL.2014.6927392

DO - 10.1109/FPL.2014.6927392

M3 - Conference contribution

AN - SCOPUS:84911165172

T3 - Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014

BT - Conference Digest - 24th International Conference on Field Programmable Logic and Applications, FPL 2014

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 24th International Conference on Field Programmable Logic and Applications, FPL 2014

Y2 - 1 September 2014 through 5 September 2014

ER -

Run-time accelerator binding for tile-based mixed-grained reconfigurable architectures

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this