WPVM: Parallel computing for the people

Alves, A.C.P.; Silva, Luís Moura; Carreira, João; Silva, João Gabriel

doi:10.1007/bfb0046684

Cited by 20 publications

(21 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For a heterogeneous virtual machine T L and T E depend on the processor speed S. Several experiments were conducted in order to measure these parameters, for the network referred to in the results section, which is composed by processors as illustrated in [1]. Therefore, to model correctly a broadcast the time spent in one message has to be multiplied by p ; 1.…”

Section: Computational Modelmentioning

confidence: 99%

“…For related grids, e.g. switching from (1,6) to (1,7), the system evaluates if the gain in time due to the addition of one processor is overcome by the data redistribution time.…”

Section: Data Redistributionmentioning

confidence: 99%

“…As an example, for a machine composed by 6 processors of capacity f244, 244, 161, 161, 60, 50g Mflops, l i would be f0.265, 0.265, 0.175, 0.175, 0.065, 0.054g. To distribute a matrix of size 1800 over a (1,6) processor grid the assignment would be 1800 rows by f477, 477, 315, 315, 118, 98g columns respectively.…”

Section: Data Distributionmentioning

confidence: 99%

“…T T (n p) = T S (n p) + T P (n p) (1) The interconnection network is modeled by a temporal expression, T C , representing the time required to transmit a message of nb bits between two network nodes, assuming a distance 1 network.…”

Section: Computational Modelmentioning

confidence: 99%

“…Existing software, such as the Windows Parallel Virtual Machine (WPVM) [1], allows building parallel virtual computers by integrating in a common processing environment a set of distinct machines (nodes) connected to the network.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Linear algebra algorithms in a heterogeneous cluster of personal computers

Barbosa

Tavares

Padilha

Proceedings 9th Heterogeneous Computing Workshop (HCW 2000) (Cat. No.PR00556)

View full text Add to dashboard Cite

show abstract

Section: Computational Modelmentioning

confidence: 99%

“…For related grids, e.g. switching from (1,6) to (1,7), the system evaluates if the gain in time due to the addition of one processor is overcome by the data redistribution time.…”

Section: Data Redistributionmentioning

confidence: 99%

Section: Data Distributionmentioning

confidence: 99%

Section: Computational Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Linear algebra algorithms in a heterogeneous cluster of personal computers

Barbosa

Tavares

Padilha

Proceedings 9th Heterogeneous Computing Workshop (HCW 2000) (Cat. No.PR00556)

View full text Add to dashboard Cite

show abstract

Heterogeneous parallel computing using Java and WMPI

Silva

Martins

Silva

2000

Concurrency: Pract. Exper.

Self Cite

View full text Add to dashboard Cite

In this paper, we present briefly the implementation of a Java interface for WMPI, a Windows-based implementation of MPI. Then, we describe a system that is oriented for Web-based computing and present a solution to integrate WMPI with this tool by making use of a Java bridge component and the Java bindings for WMPI. This solution allows the execution of meta-applications over a mixed configuration of platforms, execution models and programming languages. The resulting system provides a way to solve the problem of heterogeneity and to unleash the potential of diverse computational resources and programming tools.One of the goals of the JavaGrande Forum is to define a specification for a Java MPI API, and a position document is available at [3]. There are some implementations of Java bindings for MPI, like mpiJava [4], JavaMPI [5], MPIJ [6] and JMPI [7].Since 1994 we have been involved in the development of MPI and PVM libraries for Windows NT [8,9] and both implementations (WPVM and WMPI) have been highly used. We also provide a Java-based interface for Windows PVM and MPI. We have ported the jPVM interface [10] for WPVM, and a Java binding for WMPI has been developed from scratch. In this paper we briefly describe the implementation of the JWMPI interface.These Java bindings allow the communication with C/C++ applications by using MPI or PVM. The idea is not to replace all the software written in traditional languages with new Java programs. On the contrary, the access to standard libraries is essential not only for performance reasons, but also for software engineering considerations: it would allow existing Fortran and C code to be reused at a reduced cost when writing new applications in Java.The rest of the paper is organized as follows: the next section presents a brief overview of WPVM and WMPI libraries. Section 3 describes the features of our Java binding for WMPI, while Section 4 describes how JWMPI has been integrated with another Java-based tool that is oriented to Web-based computing. Section 5 presents some performance results. The related work is described in Section 6, while Section 7 concludes the paper. IMPLEMENTATIONS OF PVM AND MPI FOR WINDOWS NTWPVM and WMPI are full ports of the standard specifications of PVM and MPI, thereby ensuring that parallel applications developed on top of PVM and MPI can be executed in the MS Windows operating system, as long as they do not use any special feature of the underlying operating system. These ports can run in heterogeneous clusters of Windows 95/NT and Unix machines. WPVM (Windows Parallel Virtual Machine) is an implementation of the PVM message passing environment as defined in release 3.3, the original PVM package from the Oak Ridge National Laboratory. WPVM includes libraries for Borland C++ 4.51, Microsoft VisualC++ 2.0, Watcom 10.5 and Microsoft Fortran PowerStation (v1.3). The library is available at [11]. On the other hand, WMPI is an implementation of the Message Passing Interface standard for Microsoft Win32 platforms. It is based on MPICH 1.0....

show abstract

Algorithm-Dependant Method to Determine the Optimal Number of Computers in Parallel Virtual Machines

Barbosa¹,

Padilha²

1999

Vector and Parallel Processing – VECPAR’98

View full text Add to dashboard Cite

WPVM: Parallel computing for the people

Cited by 20 publications

References 2 publications

Linear algebra algorithms in a heterogeneous cluster of personal computers

Linear algebra algorithms in a heterogeneous cluster of personal computers

Heterogeneous parallel computing using Java and WMPI

Algorithm-Dependant Method to Determine the Optimal Number of Computers in Parallel Virtual Machines

Contact Info

Product

Resources

About