Realizing Efficient Execution of Dataflow Actors on Manycores

Essayas Gebrewahid; Mingkun Yang; Gustav Cedersjö; Zain Ul-Abdin; Veronica Gaspes; Jörn Janneck; Bertil Svensson

doi:10.1109/EUC.2014.55

Realizing Efficient Execution of Dataflow Actors on Manycores

Essayas Gebrewahid, Mingkun Yang, Gustav Cedersjö, Zain Ul-Abdin, Veronica Gaspes, Jörn Janneck, Bertil Svensson

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

Abstract

Embedded DSP computing is currently shifting towards manycore architectures in order to cope with the ever growing computational demands. Actor based dataflow languages are being considered as a programming model. In this paper we present a code generator for CAL, one such dataflow language. We propose to use a compilation tool with two intermediate representations. We start from a machine model of the actors that provides an ordering for testing of conditions and firing of actions. We then generate an Action Execution Intermediate Representation that is closer to a sequential imperative language like C and Java. We describe our two intermediate representations and show the feasibility and portability of our approach by compiling a CAL implementation of the Two-Dimensional Inverse Discrete Cosine Transform on a general purpose processor, on the Epiphany manycore architecture and on the Ambric massively parallel processor array.

Original language	English
Title of host publication	2014 12th IEEE International Conference on Embedded and Ubiquitous Computing (EUC 2014)
Publisher	IEEE - Institute of Electrical and Electronics Engineers Inc.
Pages	321-328
DOIs	https://doi.org/10.1109/EUC.2014.55
Publication status	Published - 2014
Event	12th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC) - Milano, ITALY Duration: 2014 Aug 6 → 2014 Aug 8

Conference

Conference	12th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC)
Period	2014/08/06 → 2014/08/08

Subject classification (UKÄ)

Computer Sciences

Free keywords

dataflow languages
compilation framework
code generation
manycore
CAL

Access to Document

10.1109/EUC.2014.55

Cite this

Gebrewahid, E., Yang, M., Cedersjö, G., Ul-Abdin, Z., Gaspes, V., Janneck, J., & Svensson, B. (2014). Realizing Efficient Execution of Dataflow Actors on Manycores. In 2014 12th IEEE International Conference on Embedded and Ubiquitous Computing (EUC 2014) (pp. 321-328). IEEE - Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/EUC.2014.55

@inproceedings{b6e479df3a394c22b176ee212f490ea7,

title = "Realizing Efficient Execution of Dataflow Actors on Manycores",

abstract = "Embedded DSP computing is currently shifting towards manycore architectures in order to cope with the ever growing computational demands. Actor based dataflow languages are being considered as a programming model. In this paper we present a code generator for CAL, one such dataflow language. We propose to use a compilation tool with two intermediate representations. We start from a machine model of the actors that provides an ordering for testing of conditions and firing of actions. We then generate an Action Execution Intermediate Representation that is closer to a sequential imperative language like C and Java. We describe our two intermediate representations and show the feasibility and portability of our approach by compiling a CAL implementation of the Two-Dimensional Inverse Discrete Cosine Transform on a general purpose processor, on the Epiphany manycore architecture and on the Ambric massively parallel processor array.",

keywords = "dataflow languages, compilation framework, code generation, manycore, CAL",

author = "Essayas Gebrewahid and Mingkun Yang and Gustav Cedersj{\"o} and Zain Ul-Abdin and Veronica Gaspes and J{\"o}rn Janneck and Bertil Svensson",

year = "2014",

doi = "10.1109/EUC.2014.55",

language = "English",

pages = "321--328",

booktitle = "2014 12th IEEE International Conference on Embedded and Ubiquitous Computing (EUC 2014)",

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

address = "United States",

note = "12th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC) ; Conference date: 06-08-2014 Through 08-08-2014",

}

Gebrewahid, E, Yang, M, Cedersjö, G, Ul-Abdin, Z, Gaspes, V, Janneck, J & Svensson, B 2014, Realizing Efficient Execution of Dataflow Actors on Manycores. in 2014 12th IEEE International Conference on Embedded and Ubiquitous Computing (EUC 2014). IEEE - Institute of Electrical and Electronics Engineers Inc., pp. 321-328, 12th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC), 2014/08/06. https://doi.org/10.1109/EUC.2014.55

Realizing Efficient Execution of Dataflow Actors on Manycores. / Gebrewahid, Essayas; Yang, Mingkun; Cedersjö, Gustav et al.
2014 12th IEEE International Conference on Embedded and Ubiquitous Computing (EUC 2014). IEEE - Institute of Electrical and Electronics Engineers Inc., 2014. p. 321-328.

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

TY - GEN

T1 - Realizing Efficient Execution of Dataflow Actors on Manycores

AU - Gebrewahid, Essayas

AU - Yang, Mingkun

AU - Cedersjö, Gustav

AU - Ul-Abdin, Zain

AU - Gaspes, Veronica

AU - Janneck, Jörn

AU - Svensson, Bertil

PY - 2014

Y1 - 2014

N2 - Embedded DSP computing is currently shifting towards manycore architectures in order to cope with the ever growing computational demands. Actor based dataflow languages are being considered as a programming model. In this paper we present a code generator for CAL, one such dataflow language. We propose to use a compilation tool with two intermediate representations. We start from a machine model of the actors that provides an ordering for testing of conditions and firing of actions. We then generate an Action Execution Intermediate Representation that is closer to a sequential imperative language like C and Java. We describe our two intermediate representations and show the feasibility and portability of our approach by compiling a CAL implementation of the Two-Dimensional Inverse Discrete Cosine Transform on a general purpose processor, on the Epiphany manycore architecture and on the Ambric massively parallel processor array.

AB - Embedded DSP computing is currently shifting towards manycore architectures in order to cope with the ever growing computational demands. Actor based dataflow languages are being considered as a programming model. In this paper we present a code generator for CAL, one such dataflow language. We propose to use a compilation tool with two intermediate representations. We start from a machine model of the actors that provides an ordering for testing of conditions and firing of actions. We then generate an Action Execution Intermediate Representation that is closer to a sequential imperative language like C and Java. We describe our two intermediate representations and show the feasibility and portability of our approach by compiling a CAL implementation of the Two-Dimensional Inverse Discrete Cosine Transform on a general purpose processor, on the Epiphany manycore architecture and on the Ambric massively parallel processor array.

KW - dataflow languages

KW - compilation framework

KW - code generation

KW - manycore

KW - CAL

U2 - 10.1109/EUC.2014.55

DO - 10.1109/EUC.2014.55

M3 - Paper in conference proceeding

SP - 321

EP - 328

BT - 2014 12th IEEE International Conference on Embedded and Ubiquitous Computing (EUC 2014)

PB - IEEE - Institute of Electrical and Electronics Engineers Inc.

T2 - 12th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC)

Y2 - 6 August 2014 through 8 August 2014

ER -

Realizing Efficient Execution of Dataflow Actors on Manycores

Abstract

Conference

Subject classification (UKÄ)

Free keywords

Access to Document

Fingerprint

Cite this