Mapping the Join Calculus to Heterogeneous Hardware

TL;DR

This paper proposes a method to map the Join Calculus, a model for concurrent computation, onto heterogeneous hardware architectures by leveraging non-determinism to unify placement and scheduling, reducing platform-specific software development.

Contribution

It introduces a novel Cartesian-product-style mapping of the Join Calculus to heterogeneous hardware, integrating placement and scheduling into a single unified task.

Findings

Demonstrates the feasibility of the mapping approach

Shows how non-determinism encodes placement choices

Unifies placement and scheduling concepts

Abstract

As modern architectures introduce additional heterogeneity and parallelism, we look for ways to deal with this that do not involve specialising software to every platform. In this paper, we take the Join Calculus, an elegant model for concurrent computation, and show how it can be mapped to an architecture by a Cartesian-product-style construction, thereby making use of the calculus' inherent non-determinism to encode placement choices. This unifies the concepts of placement and scheduling into a single task.