Coupling DSM-Based Parallel Applications
نویسنده
چکیده
When coupling applications running on distributed memory architectures or clusters, the coupling library must adapt to the distribution of the data in the memory of each computation node. The library must be prepared to redistribute the data when the coupled applications use different data mappings or when the number of processors of the two architectures are different. Mome is a user-level software DSM which allows programs running on a distributed memory architecture or cluster to create segments and to share data objects through memory mapping. The segments of the DSM form a simple linear address space where all shared objects of applications are allocated. The Mome coupling library accesses the data through mappings of the DSM segments on the memories of the communication threads. The parallel communication threads are distributed on the computation nodes and exploit the communication capacity of each processor. The data are moved directly between the DSM segments and the transfers do not rely on any knowledge on the application use of these segments. Key-words: DSM, memory mapping, code coupling, parallel communication.
منابع مشابه
Evaluation of the JIAJIA Software DSM System on High Performance Computer Architectures
Distributed Shared Memory (DSM) combines the scalability of loosely coupled multicomputer systems with the ease of usability of tightly coupled multiprocessors, and allows transparent replication and caching of data. DSM has received much attention in the past decade and many consistency models, protocols, and systems were developed. In this paper, we describe a new software DSM system called J...
متن کاملHomeless and Home-based Lazy Release Consistency Protocols on Distributed Shared Memory
This paper describes the comparison between homeless and home-based Lazy Release Consistency (LRC) protocols which are used to implement Distributed Shared Memory (DSM) in cluster computing. We present a performance evaluation of parallel applications running on homeless and home-based LRC protocols. We compared the performance between TreadMarks, which uses homeless LRC protocol, and our home-...
متن کاملRunning Real Applications on Software DSMs
This paper introduces our experiences with some real applications on the home-based software DSM JIAJIA and discusses techniques of parallelizing a sequential program to run on software DSM. It categorizes parallel program segments into ve patterns: single-process sequential, mutual-exclusive sequential, data-parallel, task-parallel, and common-parallel. The usage of each pattern is then discus...
متن کاملParleda: a Library for Parallel Processing in Computational Geometry Applications
ParLeda is a software library that provides the basic primitives needed for parallel implementation of computational geometry applications. It can also be used in implementing a parallel application that uses geometric data structures. The parallel model that we use is based on a new heterogeneous parallel model named HBSP, which is based on BSP and is introduced here. ParLeda uses two main lib...
متن کاملProgramming Large Dynamic Data Structures on a DSM Cluster of Multicores∗
Applications in increasingly important domains such as data mining and graph analysis operate on very large, dynamically constructed graphs, i.e. they are composed of dynamically allocated objects linked together via pointers. Parallel algorithms on large graphs can greatly benefit from software Distributed Shared Memory’s (DSM) convenience of sharedmemory programming and computational scalabil...
متن کامل