Shared-variable synchronization approaches for dynamic dataflow programs

Apostolos Modas; Simone Casale-Brunet; Robert James Stewart; Endri Bezati; Junaid Ahmad; Marco Mattavelli

Shared-variable synchronization approaches for dynamic dataflow programs

Apostolos Modas
, Simone Casale-Brunet
, Robert James Stewart
, Endri Bezati
, Junaid Ahmad
, Marco Mattavelli

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

91 Downloads (Pure)

Abstract

This paper presents shared-variable synchronization approaches for dataflow programming. The mechanisms do not require any substantial model of computation (MoC) modification, and is portable across both for hardware (HW) and
software (SW) low-level code synthesis. With the shared-variable
formalization, the benefits of the dataflow MoC are maintained,
however the space and energy efficiency of an application can be
significantly improved. The approach targets Dynamic Process
Network (DPN) dataflow applications, thus making them also suitable for less expressive models e.g. synchronous and cyclo-static dataflow that DPN subsumes. The approach is validated through the analysis and optimization of a High-Efficiency Video Coding (HEVC) decoder implemented in the RVC-CAL dataflow language targeting a multi-core platform. Experimental results show how, starting from an initial design that does not use
the shared-variable formalism, frames per second throughput
performance is increased by a factor of 21.

Original language	English
Title of host publication	Workshop on Signal Processing Systems 2018
Publisher	IEEE
Publication status	Published - Oct 2018
Event	IEEE Workshop on Signal Processing Systems 2018 - Cape Town, South Africa Duration: 21 Oct 2018 → 24 Oct 2018

Workshop

Workshop	IEEE Workshop on Signal Processing Systems 2018
Abbreviated title	IEEE SiPS
Country/Territory	South Africa
City	Cape Town
Period	21/10/18 → 24/10/18

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Access to Document

Download pdf
© IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Accepted author manuscript, 826 KBLicence: All Rights Reserved

Cite this

@inproceedings{0830b9f948964e468a4bce817ac5b5e8,

title = "Shared-variable synchronization approaches for dynamic dataflow programs",

abstract = "This paper presents shared-variable synchronization approaches for dataflow programming. The mechanisms do not require any substantial model of computation (MoC) modification, and is portable across both for hardware (HW) and software (SW) low-level code synthesis. With the shared-variable formalization, the benefits of the dataflow MoC are maintained, however the space and energy efficiency of an application can be significantly improved. The approach targets Dynamic Process Network (DPN) dataflow applications, thus making them also suitable for less expressive models e.g. synchronous and cyclo-static dataflow that DPN subsumes. The approach is validated through the analysis and optimization of a High-Efficiency Video Coding (HEVC) decoder implemented in the RVC-CAL dataflow language targeting a multi-core platform. Experimental results show how, starting from an initial design that does not use the shared-variable formalism, frames per second throughput performance is increased by a factor of 21.",

author = "Apostolos Modas and Simone Casale-Brunet and Stewart, \{Robert James\} and Endri Bezati and Junaid Ahmad and Marco Mattavelli",

year = "2018",

month = oct,

language = "English",

booktitle = "Workshop on Signal Processing Systems 2018",

publisher = "IEEE",

address = "United States",

note = "IEEE Workshop on Signal Processing Systems 2018, IEEE SiPS ; Conference date: 21-10-2018 Through 24-10-2018",

}

TY - GEN

T1 - Shared-variable synchronization approaches for dynamic dataflow programs

AU - Modas, Apostolos

AU - Casale-Brunet, Simone

AU - Stewart, Robert James

AU - Bezati, Endri

AU - Ahmad, Junaid

AU - Mattavelli, Marco

PY - 2018/10

Y1 - 2018/10

N2 - This paper presents shared-variable synchronization approaches for dataflow programming. The mechanisms do not require any substantial model of computation (MoC) modification, and is portable across both for hardware (HW) and software (SW) low-level code synthesis. With the shared-variable formalization, the benefits of the dataflow MoC are maintained, however the space and energy efficiency of an application can be significantly improved. The approach targets Dynamic Process Network (DPN) dataflow applications, thus making them also suitable for less expressive models e.g. synchronous and cyclo-static dataflow that DPN subsumes. The approach is validated through the analysis and optimization of a High-Efficiency Video Coding (HEVC) decoder implemented in the RVC-CAL dataflow language targeting a multi-core platform. Experimental results show how, starting from an initial design that does not use the shared-variable formalism, frames per second throughput performance is increased by a factor of 21.

AB - This paper presents shared-variable synchronization approaches for dataflow programming. The mechanisms do not require any substantial model of computation (MoC) modification, and is portable across both for hardware (HW) and software (SW) low-level code synthesis. With the shared-variable formalization, the benefits of the dataflow MoC are maintained, however the space and energy efficiency of an application can be significantly improved. The approach targets Dynamic Process Network (DPN) dataflow applications, thus making them also suitable for less expressive models e.g. synchronous and cyclo-static dataflow that DPN subsumes. The approach is validated through the analysis and optimization of a High-Efficiency Video Coding (HEVC) decoder implemented in the RVC-CAL dataflow language targeting a multi-core platform. Experimental results show how, starting from an initial design that does not use the shared-variable formalism, frames per second throughput performance is increased by a factor of 21.

M3 - Conference contribution

BT - Workshop on Signal Processing Systems 2018

PB - IEEE

T2 - IEEE Workshop on Signal Processing Systems 2018

Y2 - 21 October 2018 through 24 October 2018

ER -

Shared-variable synchronization approaches for dynamic dataflow programs

Abstract

Workshop

UN SDGs

Access to Document

Fingerprint

Cite this