Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Scalable Actor Networks with CAL

Callanan, Gareth LU and Gruian, Flavius LU orcid (2023) p.169-179
Abstract
Dataflow is a Model of Computation (MoC) that describes applications as networks of actors. The CAL Actor Language (CAL) is one of the programming languages for describing such actors. A downside to CAL is that the actors and their networks are rigidly defined - it is not possible to have a parametric number of ports or actions in an actor. This makes it difficult to define flexible applications or large actors without extra programmer effort.

The key contribution of this paper is the addition of constructs to CAL to allow for parametric actor and network creation. These constructs include defining and using arrays and sub-arrays ports, as well as a generate statement for creating multiple similar actions. We give examples of the... (More)
Dataflow is a Model of Computation (MoC) that describes applications as networks of actors. The CAL Actor Language (CAL) is one of the programming languages for describing such actors. A downside to CAL is that the actors and their networks are rigidly defined - it is not possible to have a parametric number of ports or actions in an actor. This makes it difficult to define flexible applications or large actors without extra programmer effort.

The key contribution of this paper is the addition of constructs to CAL to allow for parametric actor and network creation. These constructs include defining and using arrays and sub-arrays ports, as well as a generate statement for creating multiple similar actions. We give examples of the types of parametric applications that our new constructs enable.

We use our work to demonstrate some of the limitations in dataflow tools when applied to non-trivial actors. We focus on the processing effort required to transform an actor into an intermediate representation (IR). We make the case for future research to use parametric actors to test the scaling of dataflow toolchains.
(Less)
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
MEMOCODE '23: 21st ACM-IEEE International Conference on Formal Methods and Models for System Design
pages
169 - 179
publisher
Association for Computing Machinery (ACM)
external identifiers
  • scopus:85179624377
ISBN
979-8-4007-0318-8/23/09
DOI
10.1145/3610579.3611074
project
Investigation on Describing, Compiling and Analysing Scalable Dataflow Actor Networks in Streaming Applications
language
English
LU publication?
yes
id
94d61405-7593-4d92-a6d8-74f4ecd7c22e
date added to LUP
2023-12-11 10:03:55
date last changed
2024-01-11 15:04:49
@inproceedings{94d61405-7593-4d92-a6d8-74f4ecd7c22e,
  abstract     = {{Dataflow is a Model of Computation (MoC) that describes applications as networks of actors. The CAL Actor Language (CAL) is one of the programming languages for describing such actors. A downside to CAL is that the actors and their networks are rigidly defined - it is not possible to have a parametric number of ports or actions in an actor. This makes it difficult to define flexible applications or large actors without extra programmer effort.<br/><br/>The key contribution of this paper is the addition of constructs to CAL to allow for parametric actor and network creation. These constructs include defining and using arrays and sub-arrays ports, as well as a generate statement for creating multiple similar actions. We give examples of the types of parametric applications that our new constructs enable.<br/><br/>We use our work to demonstrate some of the limitations in dataflow tools when applied to non-trivial actors. We focus on the processing effort required to transform an actor into an intermediate representation (IR). We make the case for future research to use parametric actors to test the scaling of dataflow toolchains.<br/>}},
  author       = {{Callanan, Gareth and Gruian, Flavius}},
  booktitle    = {{MEMOCODE '23: 21st ACM-IEEE International Conference on Formal Methods and Models for System Design}},
  isbn         = {{979-8-4007-0318-8/23/09}},
  language     = {{eng}},
  month        = {{12}},
  pages        = {{169--179}},
  publisher    = {{Association for Computing Machinery (ACM)}},
  title        = {{Scalable Actor Networks with CAL}},
  url          = {{http://dx.doi.org/10.1145/3610579.3611074}},
  doi          = {{10.1145/3610579.3611074}},
  year         = {{2023}},
}