"Plug and Play" Parallel Computing Combines Rendezvous, Multiprocessing, and Cluster Computing
Huntington Beach, California, USA - October 16, 2002 - Dauger Research, Inc., has shipped version 1.3 of Pooch, the Parallel OperatiOn and Control Heuristic application. The latest version of the cluster management software now enables application writers to take advantage of multiprocessor Macs and cluster computing using just one API.
Pooch, given the newest "most innovative" award by IEEE Cluster, is designed to combine powerful, numerically-intensive parallel-computing clusters with the famed ease-of-use of the Macintosh.
It provides the user interface for the latest incarnation of
AppleSeed, a project begun by
physics researchers at UCLA in 1998. For four years and counting, their software is being used world-wide to transform Macintoshes into easy-to-use, numerically-intensive parallel computers.
The latest version can now launch multiple parallel-computing tasks per node, automatically taking advantage of multiprocessor Macs and OS X's preemptive multitasking.
Applications can therefore take advantage of parallel computing both across nodes and inside nodes simultaneously using only one API:
Parallel applications already written using MPI need no modification to utilize multiprocessor Macs; the extra processor merely appears to be another node. With MPI and Pooch, applications can both use multiprocessing and get "outside the box".
In addition, Pooch now uses
(a. k. a.
ZeroConf), supplementing Service Location Protocol,
to perform registration, discovery, and resolution of cluster nodes over local networks and the Internet. While Rendezvous makes for the fastest discovery of nodes yet,
Pooch retains SLP for compatibility with OS 9 and previous OS X's.
This new feature places Pooch among the latest innovative products to take advantage of new
technologies introduced in OS X Jaguar (10.2).
Pooch builds these new features on its unsurpassed capabilities, such as its "only seconds" installation, patent-pending technologies, "Computing Grid"-like automation features,
its ability to combine nodes over the Internet, customizable job queuing through
AppleScript, 512-bit command encryption, field-tested 76-node scalability, its four user interfaces, cluster access for mainstream applications, support for launching the widest variety of application types and languages (see the SDK), dynamic responsiveness to cluster conditions, support of three different Message-Passing Interfaces, self-scheduling node registration, and independence of File Sharing, NFS, rsh, and static cluster data. Pooch's combination of flexibility, simplicity, and capability remains unique throughout the industry.
Updates of Pooch to version 1.3 are shipping to current users with an active subscription. We highly recommend that our customers upgrade to OS X 10.2.1 for its bug fixes and performance enhancements. (Apple has fixed all of the remaining bugs present in previous versions of OS X known to significantly affect Pooch.)
Pooch v1.3 is available for US$175 for the first compute node then US$125 for each node thereafter. Check the
Pooch web site for special academic pricing.
We encourage you to revisit the Dauger Research web site.
The site features a
download version of Pooch,
new parallel applications,
a new list of
featured users, and
With a brand-new
Pooch Software Development Kit,
a discussion mailing list devoted to parallel computing, and
information about compiling and running code in parallel,
the site makes it easier than ever to write, develop, and run your own parallel code today.
Pooch requires networked Macintoshes and/or XServe's running OS 9 with CarbonLib 1.2 or later, OS X 10.2 or later, and/or OS X Server 10.2 or later with 4 MB of available RAM and 2 MB of disk space.
Dauger Research, Inc., was incorporated and founded by Dr. Dean E. Dauger. Dr. Dauger is the award-winning author of Atom in a Box and Fresnel Diffraction Explorer and has co-authored the award-winning Kai's Power Tools software from 1992 to 1994. After completing his Ph. D. in physics, he founded Dauger Research, Inc., to bridge the divides between the scientifically and technically complex and the mainstream by making high-performance computation and visualization easy to use and accessible to users.