An Architecture of Stampi: MPI Library on a Cluster of Parallel Computers

Abstract

Abstract. In this paper, we present a communication library which extends an MPI application on a single parallel machine to a cluster of parallel machines. Stampi provides some functionality which are required for constructing distributed applications and environments based on the MPI2 standard with a focus on dynamic process management. Since the mechanism of communication bridge is transparent for users, it is very useful to assemble and link MPI applications on meta-computer systems. Furthermore Stampi supports novel functions; one is the communication between a Java applet to the backend parallel computer. Another is supporting remote file-IO. Both give us a framework of distributed resource management based on an MPI communication infrastructure. This paper covers the architecture of Stampi.

Hierarchy-Aware Message-Passing in the Upcoming Many-Core Era

“…For that reason their application area is also referred to as a so-called meta-computing environment. The most common Grid-enabled MPI libraries are MPICH-G2 (Karonis et al, 2003), PACX-MPI (Gabriel et al, 1998), GridMPI (Matsuda et al, 2004), StaMPI (Imamura et al, 2000), MPICH/Madeleine (Aumage et al, 2001) and MetaMPICH (Pöppe et al, 2003), which are all proven to run large-scale applications in distributed environments. Although these meta-MPI implementations usually use native MPI support for site-internal communication, as for example provided by a site-local vendor MPI, they must also be based on at least a transport layer being capable of wide area communication for bridging and forwarding messages also to the remote sites.…”

Ocean Model Coupling with Atmospheric Model for Prediction of Radioactivity Contaminant Dispersion in Emergency

“…16) Stampi, which is based on the MPI2 specification, realizes dynamic process creation to different machines and communication between spawned ones within the scope of MPI semantics. Stampi distinguishes between internal and external communication and then automatically changes to suitable communication protocols.…”

Fault-Tolerant Mechanism of Both Job Execution and File Transfer for Integrated Nuclear Energy Simulation

“…These tools are easily used by researchers, but applications of these tools on integrated simulations are limited. However, remote procedure calls (RPC) 6) or grid-enabled MPI, such as STAMPI, 7) MPICH-G, 8) or PACX MPI, 9) can be applied to different integrated simulations, although substantial modifications of codes are required. Therefore, it is necessary to develop a grid computing technology that enables researchers in nuclear energy field to more easily operate a variety of integrated simulations.…”