Actor induction and meta-evaluation

Hewitt, Carl; Bishop, Peter; Greif, Irene; Smith, Brian Cantwell; Matson, Todd; Steiger, Richard N. de

doi:10.1145/512927.512942

Cited by 36 publications

(19 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We initially feared the loss of traditional threads could be a great weakness, but the team placed slightly higher value on correct concurrency than easy concurrency, leading to the design point illustrated here. To recover some flexibility, we are currently adding actor concurrency, using isolated and immutable permissions for safe message passing [22,24].…”

Section: Evolving a Type Systemmentioning

confidence: 99%

“…Second, sometimes communication among tasks is not required for correctness, but offers substantial performance benefits: for example, in a parallelized search algorithm, broadcasting the best-known result thus far can help all threads prune the search space. Currently unstrict code is used for a few instances of this, which motivates current work to add actors to the language [22,24].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Uniqueness and reference immutability for safe parallelism

et al. 2012

View full text Add to dashboard Cite

A key challenge for concurrent programming is that sideeffects (memory operations) in one thread can affect the behavior of another thread. In this paper, we present a type system to restrict the updates to memory to prevent these unintended side-effects. We provide a novel combination of immutable and unique (isolated) types that ensures safe parallelism (race freedom and deterministic execution). The type system includes support for polymorphism over type qualifiers, and can easily create cycles of immutable objects. Key to the system's flexibility is the ability to recover immutable or externally unique references after violating uniqueness without any explicit alias tracking. Our type system models a prototype extension to C# that is in active use by a Microsoft team. We describe their experiences building large systems with this extension. We prove the soundness of the type system by an embedding into a program logic.

show abstract

Section: Evolving a Type Systemmentioning

confidence: 99%

mentioning

confidence: 99%

Uniqueness and reference immutability for safe parallelism

et al. 2012

View full text Add to dashboard Cite

show abstract

“…The design of active objects is based on a distributed object computational model termed the Actor model [21]. An actor performs a combination of the following actions in response to incoming messages: the creation of new actors; messages sent to other actors; and updating of its state.…”

Section: Computational Modelmentioning

confidence: 99%

Optimisation strategies for distributed computing using an adaptive randomised structured network

Fung

2008

2008 International Conference on Machine Learning and Cybernetics

View full text Add to dashboard Cite

Abstract:One way to improve computational efficiency for complex engineering applications is to utilise distributed computing. In such distributed system, accessing objects through location-independent names can improve the system's transparency, scalability and reliability. Names however need to be resolved prior to passing the messages between the objects. This paper reports an Adaptive RandoMised Structured search network termed ARMS, which utilises a distributed Ant Colony Optimisation algorithms (ACO) to improve the efficiency of searching in a distributed environment. The paper further investigates different kinds of optimisation strategies in order to improve search efficiency. Simulation studies have shown ARMS is superior to Chord, a well-known structured network, under various performance measures.

show abstract

“…The design of active objects is based on a distributed object computational model termed Actor model [22]. An actor performs a combination of the following actions in response to incoming messages.…”

Section: Computational Modelmentioning

confidence: 99%

Intelligent Object Locality Naming Model in an Object-Based Distributed System for Engineering Applications

Fung

Depickere

2007

2007 International Conference on Machine Learning and Cybernetics

View full text Add to dashboard Cite

Abstract:Complex engineering applications demand powerful computing resources. An object based distributed system is suitable for such applications due to the inherent parallelism nature in the object based computing model. As the model is based on message passing, which in turns relies on location transparent object names, an efficient name translation system is therefore required to map the object names to the corresponding physical addresses dynamically. In this paper, we propose a heuristic distributed search algorithm, the hierarchy random walker (HRW), to take advantage of the locality of object communications thereby increasing the overall system performance. Analysis of the simulation results from a number of benchmark programs have demonstrated that the benefits of the proposed algorithm are in terms of flexibility and efficient to locate the most used objects. The approach will be applied to intelligent techniques for engineering problems.

show abstract

Actor induction and meta-evaluation

Cited by 36 publications

References 18 publications

Uniqueness and reference immutability for safe parallelism

Uniqueness and reference immutability for safe parallelism

Optimisation strategies for distributed computing using an adaptive randomised structured network

Intelligent Object Locality Naming Model in an Object-Based Distributed System for Engineering Applications

Contact Info

Product

Resources

About