Distributed Shared Memory Using Re ective Memory: The LAM System
نویسندگان
چکیده
In the past, acceptance of Distributed Shared Memory (DSM) systems suuered because of poor performance. Performance has primarily been limited by the availability of an interconnect media whose properties are similar to those required by a DSM system. In an eeort to compensate for inappropriate interfaces, sig-niicant research has been devoted to maximizing the utilization of the available interconnect. A class of interconnect media is available that begins to address the inhibitors to DSM performance. These interconnect media have one essential property; the interface is mapped into the address space of the process participating in the DSM system. This project focuses on the development of the LAM DSM system. LAM is a well-balanced hybrid DSM system implemented using Reeective Memory interconnect hardware with a software consistency policy and interface. A distinctive property of this system is that access time is directly proportional to the number of bytes accessed and unrelated to the size of the data structure being accessed. The system architecture, interface and results are described. Additionally, a case is made for the inherent advantage a memory-mapped interconnect has over an interconnect accessed through the operating system and network protocols.
منابع مشابه
Speci cation-Based Parameter-Model Interaction: Towards a Correct Re ection of Memory Characteristics in a DSM Cluster Simulation
A DSM (Distributed-Shared Memory) cluster is an attractive parallel computing platform for scienti c research as it provides programming advantages within a scalable and cost-e ective hardware solution. This bene t derives from the fact that a DSM system provides a shared-memory abstraction on top of a distributed-memory machine by caching data replicas locally. In this respect, a coherence pro...
متن کاملProgramming Research Group PRACTICAL BARRIER SYNCHRONISATION
We investigate the performance of barrier synchronisation on both shared-memory and distributed-memory architectures, using a wide range of techniques. The performance results obtained show that distributed-memory architectures behave predictably, although their performance for barrier synchronisation is relatively poor. For shared-memory architectures, a much larger range of implementation tec...
متن کاملConcord: Re-Thinking the Division of Labor in a Distributed Shared Memory System
A distributed shared memory system provides the abstraction of a shared address space on either a network of workstations or a distributed-memory multiprocessor. Although a distributed shared memory system can improve performance by relaxing the memory consistency model and maintaining memory coherence at a granularity speci ed by the programmer, the challenge is to o er ease of programming whi...
متن کاملStatic Optimization of Distributed Tuplespace Message Communications
Distributed memory parallel systems such as work station clusters stubbornly remain in need of software systems that provide programmers with an e ective but uncomplicated means of realizing improved ap plication performance Distributed shared memory systems are maturing to ll this need Tuplespace is a structured distributed shared memory that em bodies the generative communication model This p...
متن کاملPerformance Evaluation of View-Oriented Parallel Programming on Cluster of Computers
View-Oriented Parallel Programming(VOPP) is a novel programming style based on Distributed Shared Memory, which is friendly and easy for programmers to use. In this paper we compare VOPP with two other systems for parallel programming on clusters: LAM/MPI, a message passing system, and TreadMarks, a software distributed shared memory system. We present results for ten applications implemented a...
متن کامل