Architectural Issues and Preliminary Benchmarking of a Low-cost Network of Workstations based on Active Messages

نویسندگان

  • Giovanni Chiola
  • Giuseppe Ciaccio
چکیده

The cheapest platforms for parallel processing are Networks of Workstations (NOWs) equipped with oo-the-shelf communication hardware and high-end Personal Computers running an operating system with networking capabilities. However standard network protocols and mechanisms cannot deliver a satisfactory amount of the raw communication performance to the application level. The Genoa Active Message MAchine (GAMMA) overcomes such a limitation by implementing the Active Message communication paradigm in a NOW architecture. The Active Message layer of GAMMA drastically shortens the communication path between two communicating processes while preserving a high-level and easy-to-use interface to SPMD/MIMD programming. A Parallel Matrix Multiplication (PMM) algorithm was run on GAMMA as a preliminary benchmark. The comparison between the performance achieved by PMM on GAMMA and the performance achieved by PMM with Linux TCP/IP communications on the same hardware platform demonstrate the success of the GAMMA approach.

منابع مشابه

Implementing a Low Cost , Low Latency

The cost of high-performance parallel platforms prevents parallel processing techniques from spreading in present applications. Networks of Workstations (NOW) exploiting oo-the-shelf communication hardware, high-end PCs and standard communication software provide much cheaper but poorly performing parallel platforms. In our NOW prototype called GAMMA (Genoa Active Message MAchine) every node is...

متن کامل

GAMMA: A low-cost network of workstations based on active messages

Networks Of Workstations (NOW) are an emerging architecture capable of supporting parallel processing with signiicantly low cost/performance ratio. At the moment the implementation of standard high-level communication mechanisms in a NOW does not provide such a satisfactory cost/performance ratio, as modern communication hardware would allow. We show how a standard, Unix-like operating system k...

متن کامل

Design and Control of a Novel hybrid Active Power Filter based on Load Harmonic Currents Separation with the aim of DC-link voltage minimization in voltage source converter

This paper proposes a new hybrid active power filter, including voltage source converter (VSC) based active power filter and Thyristor Controllable LC Passive Filter (TCLC-HAPF) for eliminating harmonic load current components and compensating reactive power. This new structure which is based on the separation of high and low-frequency components of load current idea can reduce the drawbacks of...

متن کامل

Implementing a Low Cost, Low Latency Parallel Platform

The practical application of parallel processing techniques is in most cases hampered by the cost of high-performance parallel platforms. The use of a Network of Workstations (NOW) connected by a fast Local Area Network (LAN) is very appealing for implementing a low-cost parallel platform. The use of higher-end PCs connected by fast LAN and running Unix-like operating systems is even more appea...

متن کامل

An Efficiency Measurement and Benchmarking Model Based on Tobit Regression, GANN-DEA and PSOGA

The purpose of this study is designing a model based on Tobit regression, DEA, Artificial Neural Network, Genetic Algorithm and Particle Swarm Optimization to evaluate the efficiency and also benchmarking the efficient and inefficient units. This model has three stages, and it uses the data envelopment analysis combined model with neural network, optimized by genetic algorithm, to evaluate the ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

متن کامل
عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1997