Adaptive Techniques for Home-Based Software DSMs
نویسندگان
چکیده
This work introduces the HAP system, a software DSM that dynamically adapts to the parallel application’s sharing patterns. Adaptation is based on a constantly categorization of the sharing experienced by each page. Migratory and producer/consumer pages are managed in a single-writer mode and may be updated, while falsely-shared pages are managed in multiple-writer mode and under invalidated protocol. A home based protocol can potentially benefit from detecting the migratory sharing pattern and dynamically decide the best assignment of the home according with the sharing mode of the page. We performed experiments with five parallel applications on a 8-node cluster system using VI messaging passing architecture. The results show that HAP substantially improves the performance of single-writer applications with sharing pages modified in critical sections protected by lock, and that can potentially improve the performance of applications with producer/consumer and single-writer between barriers sharing patterns, but it is needed first the detection and adaptation techniques for those patterns be redesigned to show the real gains of performance that is possible and was indicated in the experiments. Keywords— Software Distributed Shared-Memory, Home-based Lazy Release Consistency, Home-based Adaptive Protocol
منابع مشابه
Adaptive Write Detection in Home-based Software DSMs
Write detection is essential in multiple-writer protocols to identify writes to shared pages so that these writes can be correctly propagated. Software DSMs that implement multiple-writer protocol normally employ the virtual memory page fault to detect writes to shared pages. It write-protects shared pages at the beginning of an interval to detect writes of the interval. This paper proposes a n...
متن کاملPerformance Optimization of Software Distributed Shared Memory Systems
Software Distributed Shared Memory Systems (DSMs, or Shared Virtual Memory) are advocated to be an ideal vehicle for parallel programming because of its combination of programmability of shared memory and scalability of distributed memory systems. The challenge in building a software DSM system is to achieve good performance over a wide range of parallel programs without requiring programmers t...
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 کاملDynamic Computation Scheduling for Load Balancing in Home-Based Software DSMs
Load balancing is a critical issue for achieving good performance in parallel and distributed systems. However , this issue is neglected in the research area of software DSMs in the past decade. In this paper, we present and evaluate a dynamic computation scheduling scheme for load balancing of iterative applications in software DSM system. The experiment platform is a home based DSM system nam...
متن کامل