Transport Triggered Array Processor for Vision Applications

Safarpour Mehdi (Corresponding author), Hautala Ilkka, Miguel Bordallo López, Olli Silvén

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

Abstract

Low-level sensory data processing in many Internet-of-Things (IoT) devices pursue energy efficiency by utilizing sleep modes or slowing the clocking to the minimum. To curb the share of stand-by power dissipation in those designs, ultra-low-leakage processes are employed in fabrication. Those limit the clocking rates significantly, reducing the computing throughputs of individual cores. In this contribution we explore compensating for the substantial computing power needs of a vision application using massive parallelism. The Processing Elements (PE) of the design are based on Transport Triggered Architecture. The fine grained programmable parallel solution allows for fast and efficient computation of learnable low-level features (e.g. local binary descriptors and convolutions). Other operations, including Max-pooling have also been implemented. The programmable design achieves excellent energy efficiency for Local Binary Patterns computations.
Original languageEnglish
Title of host publicationEmbedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019
EditorsMaxime Pelcat, Matthias Jung, Dionisios N. Pnevmatikatos
PublisherSpringer
Pages361-372
Number of pages12
ISBN (Electronic)978-3-030-27562-4
ISBN (Print)978-3-030-27561-7
DOIs
Publication statusPublished - 2019
MoE publication typeA4 Article in a conference publication
Event19th International Conference on Embedded Computer Systems, SAmos 2019 - Samos, Greece
Duration: 7 Jul 201911 Jul 2019

Publication series

SeriesLecture Notes in Computer Science
Volume11733
ISSN0302-9743

Conference

Conference19th International Conference on Embedded Computer Systems, SAmos 2019
CountryGreece
CitySamos
Period7/07/1911/07/19

Fingerprint

Parallel processing systems
Energy efficiency
Curbs
Convolution
Energy dissipation
Throughput
Fabrication
Processing

Keywords

  • Binary Patterns
  • Computer vision
  • Embedded systems
  • Internet-of-Things
  • Massive processing arrays

Cite this

Mehdi, S., Ilkka, H., Bordallo López, M., & Silvén, O. (2019). Transport Triggered Array Processor for Vision Applications. In M. Pelcat, M. Jung, & D. N. Pnevmatikatos (Eds.), Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019 (pp. 361-372). Springer. Lecture Notes in Computer Science, Vol.. 11733 https://doi.org/10.1007/978-3-030-27562-4_26
Mehdi, Safarpour ; Ilkka, Hautala ; Bordallo López, Miguel ; Silvén, Olli. / Transport Triggered Array Processor for Vision Applications. Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019. editor / Maxime Pelcat ; Matthias Jung ; Dionisios N. Pnevmatikatos. Springer, 2019. pp. 361-372 (Lecture Notes in Computer Science, Vol. 11733).
@inproceedings{a969a97ed0be4418b0af9e97daf78299,
title = "Transport Triggered Array Processor for Vision Applications",
abstract = "Low-level sensory data processing in many Internet-of-Things (IoT) devices pursue energy efficiency by utilizing sleep modes or slowing the clocking to the minimum. To curb the share of stand-by power dissipation in those designs, ultra-low-leakage processes are employed in fabrication. Those limit the clocking rates significantly, reducing the computing throughputs of individual cores. In this contribution we explore compensating for the substantial computing power needs of a vision application using massive parallelism. The Processing Elements (PE) of the design are based on Transport Triggered Architecture. The fine grained programmable parallel solution allows for fast and efficient computation of learnable low-level features (e.g. local binary descriptors and convolutions). Other operations, including Max-pooling have also been implemented. The programmable design achieves excellent energy efficiency for Local Binary Patterns computations.",
keywords = "Binary Patterns, Computer vision, Embedded systems, Internet-of-Things, Massive processing arrays",
author = "Safarpour Mehdi and Hautala Ilkka and {Bordallo L{\'o}pez}, Miguel and Olli Silv{\'e}n",
year = "2019",
doi = "10.1007/978-3-030-27562-4_26",
language = "English",
isbn = "978-3-030-27561-7",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "361--372",
editor = "Maxime Pelcat and Matthias Jung and Pnevmatikatos, {Dionisios N.}",
booktitle = "Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019",
address = "Germany",

}

Mehdi, S, Ilkka, H, Bordallo López, M & Silvén, O 2019, Transport Triggered Array Processor for Vision Applications. in M Pelcat, M Jung & DN Pnevmatikatos (eds), Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019. Springer, Lecture Notes in Computer Science, vol. 11733, pp. 361-372, 19th International Conference on Embedded Computer Systems, SAmos 2019, Samos, Greece, 7/07/19. https://doi.org/10.1007/978-3-030-27562-4_26

Transport Triggered Array Processor for Vision Applications. / Mehdi, Safarpour (Corresponding author); Ilkka, Hautala; Bordallo López, Miguel; Silvén, Olli.

Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019. ed. / Maxime Pelcat; Matthias Jung; Dionisios N. Pnevmatikatos. Springer, 2019. p. 361-372 (Lecture Notes in Computer Science, Vol. 11733).

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

TY - GEN

T1 - Transport Triggered Array Processor for Vision Applications

AU - Mehdi, Safarpour

AU - Ilkka, Hautala

AU - Bordallo López, Miguel

AU - Silvén, Olli

PY - 2019

Y1 - 2019

N2 - Low-level sensory data processing in many Internet-of-Things (IoT) devices pursue energy efficiency by utilizing sleep modes or slowing the clocking to the minimum. To curb the share of stand-by power dissipation in those designs, ultra-low-leakage processes are employed in fabrication. Those limit the clocking rates significantly, reducing the computing throughputs of individual cores. In this contribution we explore compensating for the substantial computing power needs of a vision application using massive parallelism. The Processing Elements (PE) of the design are based on Transport Triggered Architecture. The fine grained programmable parallel solution allows for fast and efficient computation of learnable low-level features (e.g. local binary descriptors and convolutions). Other operations, including Max-pooling have also been implemented. The programmable design achieves excellent energy efficiency for Local Binary Patterns computations.

AB - Low-level sensory data processing in many Internet-of-Things (IoT) devices pursue energy efficiency by utilizing sleep modes or slowing the clocking to the minimum. To curb the share of stand-by power dissipation in those designs, ultra-low-leakage processes are employed in fabrication. Those limit the clocking rates significantly, reducing the computing throughputs of individual cores. In this contribution we explore compensating for the substantial computing power needs of a vision application using massive parallelism. The Processing Elements (PE) of the design are based on Transport Triggered Architecture. The fine grained programmable parallel solution allows for fast and efficient computation of learnable low-level features (e.g. local binary descriptors and convolutions). Other operations, including Max-pooling have also been implemented. The programmable design achieves excellent energy efficiency for Local Binary Patterns computations.

KW - Binary Patterns

KW - Computer vision

KW - Embedded systems

KW - Internet-of-Things

KW - Massive processing arrays

UR - https://arxiv.org/abs/1906.04258v1

UR - http://www.scopus.com/inward/record.url?scp=85071473163&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-27562-4_26

DO - 10.1007/978-3-030-27562-4_26

M3 - Conference article in proceedings

SN - 978-3-030-27561-7

T3 - Lecture Notes in Computer Science

SP - 361

EP - 372

BT - Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019

A2 - Pelcat, Maxime

A2 - Jung, Matthias

A2 - Pnevmatikatos, Dionisios N.

PB - Springer

ER -

Mehdi S, Ilkka H, Bordallo López M, Silvén O. Transport Triggered Array Processor for Vision Applications. In Pelcat M, Jung M, Pnevmatikatos DN, editors, Embedded Computer Systems: Architectures, Modeling, and Simulation. SAMOS 2019. Springer. 2019. p. 361-372. (Lecture Notes in Computer Science, Vol. 11733). https://doi.org/10.1007/978-3-030-27562-4_26