Sunday, November 27, 2005

Ninf-G roll on Rocks 4.1 (x86, x86_64) is Released.


Ninf-G is developed by the National Institute of Advanced Industrial Science and Technology (AIST) and National Institute of Informatics (NII) in Japan. The Ninf-G official website can be found at ninf.apgrid.org, and includes complete user and programmer documentation in both English and Japanese. Ninf-G is a Remote Proceduce Call (RPC) implementation for use on Globus-based Grids. The API implemented within Ninf-G is a reference implementation of the proposed GGF GridRPC API standard. GridRPC adapts effectively to wide variety of applications and it is especially suitable for developing general and scalable task-parallel applications. Ninf-G includes both C and Java APIs for client programming, an Interface Description Language (IDL) compiler for stub generation, and several commands for system management and utilization.

Ninf-G originated from the Ninf system, released by Ninf project at the Electrotechnical Laboratory in 1996. Development of Ninf-G was started in 2000 and it has been designed to work with basic Grid services, such as GSI, Pre-WS GRAM, and MDS in the Globus Toolkit. Ninf-G provides a simple, yet powerful, client-server-based standard RPC mechanism, plus asynchronous, coarse-grained parallel tasking frameworks for programming on a Grid, while hiding the complexity of using these services from application programmers. Ninf-G is currently used as key middleware several Grid projects including: Pacific Rim Applications and Grid Middleware Assembly (PRAGMA), Asia Pacific Partnership for Grid Computing (ApGrid), Application developments in TeraGrid, and Grid Superscaler.

Put your Mammo (graph) on the Grid

Medical imaging is one of the biggest growth areas for deploying both grid and on-demand technologies. Take the University of Pennsylvania's National Digital Mammography Archive (NDMA).
The project, which involves medical imaging company i3 and NDMA, launched three years ago with IBM at the same time Big Blue laid out its On Demand computing strategy.
When it began, the NDMA project counted four hospitals contributing mammography images. Today, some 24 hospitals are accessing the database of more than 1 million digital mammography images, which provides access to records that help about 300 doctors and researchers in their diagnosis and treatment of breast cancer.
Now, that archive of some 350,000 mammography records is poised to expand even more with a portal offering that lets women use the database as a secure, digital "locker" where they can store their own mammography images and other medical records.

For $9.95 a year (down from the usual fee of $14.95 a year), women can rent space on the grid, in a hosted, secure environment provided by IBM on the MyNDMA.com portal space offering.
"That made it impossible to have on-demand access from another hospital or radiology center. As a result, diagnostics tests and additional mammograms may be required whenever a woman visits a new doctor."

This way, he added, women can store their personal mammography images and, in the process, proactively monitor their personal health over time. "If a woman's been diagnosed with breast cancer, she typically might see several different doctors through the course of her treatment," said Danois. Now, "they can easily access records when visiting a new doctor or obtaining a second opinion."

If they want their data used in ongoing breast cancer research, that's their call. They can opt-in or opt-out when they sign up.

The servers started out with IBM's Intel based, 8-way x-Series configured for a 12-terabyte grid system. Today, it runs as a 32 node, 64-CPU system capable of handling 100 Terabytes of data. By next year, that will have expanded to a 64 node, 128 CPU, 1000 Terabyte grid system, IBM said.

Saturday, November 26, 2005

First OSS Physics Book that I have seen

THE PHYSICS TEXTBOOK by Christoph Schiller

I visited this site and I was greeted with;
How does one empty a bottle as rapidly as possible?
What are the highest force and power in nature?
How does one connect water pipes to a turning wheel?
What are the dangers of a can of beans?
What is the single page of unsolved problems in fundamental physics?

I was hooked, The site remained on a firefox tab, until I decided to download the book.

Christoph Schiller, at motion mountain has done it. I may have told before that my first love is Physics, but I was not the kind, Christoph is. I prefer to enlarge my knowledge while he tries to enlarge to knowledge of the world, who ever interested in physics. I think this is an achievement that I could never accomplish with all the grids, clusters, high performance computing or my physics knowledge with the present track that I am following.
Christoph has the whole book online and one could download the complete book or individual chapters that interest one. Either way anyone who is slightly interested in Physics, will bookmark this site once one visits.

The text is free - under two conditions: that (1) you do not charge money to others for anything containing any part of the text; and (2) you send a feedback email to fb @ motion mountain . net (without the spaces) with your comments on the sections you looked at. In return, I will answer every email with questions or suggestions. For an email with good questions or suggestions you will be mentioned in the acknowledgments or receive a reward - or both.

Some BIO on the author;
Christoph Schiller was born in 1960 and raised in Varese.He studied physics at the Universitt Stuttgart and received his Ph.D. in physics at the Universit© Libre de Bruxelles, in the department of Ilya Prigogine. He conceived and started this project in Yokohama, then wrote most of the text in Eindhoven.Other parts were added in Berlin, Mnchen, Paris and Varese.He regularly gives talks on themes from the text.

Happy reading.

Friday, November 25, 2005

The Grid Computing Weblog: Grid Standards Groups Weigh Merger

This could be good news for all those who are interested in Grid technology. Fond while browsing a fello blogger.
The Grid Computing Weblog: Grid Standards Groups Weigh Merger

Wednesday, November 23, 2005

SETI@home Will be Killed On Dec 15th, Will get a new life at BOINC

The well known Distributed computing experiment SETI@home will be switched off on December 15. Don't get alarmed, it will still be actiove under the BOINC (Berkeley Open Infrastructure for Network Computing) and continue to do the work. So if you are a member, you should have gotten an email by now with instructions as how to rejoin the project under BOINC. I know the inquirer reported as "SETI@home killed off ". They always try to be dramatic, and their acting is quite good.
I am a member of two BOINC projects;
Einstein@home: search for gravitational signals emitted by pulsars
World Community Grid: advance our knowledge of human disease. (Requires 5.2.1 or greater)
So Come join us, Find an alien, no not that type, Extraterrestrial Intelligence type!

Below are some parts of the technical news leading to the switch over;
November 22, 2005 - 21:30 UTC
We began sending out the mass e-mail yesterday warning SETI@home classic users that we are going to close down the old project on December 15th. It was sent to all 200,000 of the active classic users by this morning. Inactive classic users are being e-mailed at this point.

Due to the influx of new BOINC users (and the unfortunate timing of some googlebots and other web spiders) the load on our web server was extremely high for the past 12 hours. To fix this, we finally deployed a second web server to split the load. As DNS updates spread throughout the internet, the load on klaatu (the original single web server) decreases while the load on penguin (the new secondary web server) increases. Both are Sun D220R's (2 x 440MHz Sparc, 2 GB RAM).

Part of the problem was that the web servers were configured to spawn more many clients than actually necessary, which left lingering, unused threads open on the database, which in turn lead to the database running out of connections. Some users saw messages to this effect when the load on the web servers was at its highest.

This looked like a database problem, when in fact we are currently enjoying a 10% performance boost on the database. Last week we moved some memory off the myisam tables (which contain web forum info and not much else) and slated it for the innodb tables (which contain user, host, result, workunit, etc. tables). The myisam tables didn't need the excess memory.

By the way, the master database merge (see below) is currently on for the beginning of next week.

November 15, 2005 - 21:00 UTC
(updated 22:45 UTC - see addendum below)

Today we started the big master database merge. This step is simple in essence: we are combining all the scientific data from SETI@home classic and SETI@home/BOINC into one big database. However, this is the culmination of many months of effort.

What happened during those months? Among other things, we had to migrate all the data off of one server onto another, find and remove redundant data, add new fields to old records and populate them, write and test software to merge databases while keeping all relational constraints intact... Basically a lot of cleanup, a lot of testing, and backing up the entire set of databases between every major step.
October 10, 2005 - 23:30 UTC
Last week we finished the first phase of the science "migration." What this means is that all of the scientific results from SETI@home classic have been migrated over to the BOINC science database. There is a bit of cleanup work to be done, but the next big phase is merging these data with BOINC data. We will soon be able to shut down the classic master science database, which is currently running on the same server as the BOINC scheduler.

Tor: An anonymous Internet communication system 0.1.1.9-alpha ready for download

Tor is a toolset for a wide range of organizations and people that want to improve their safety and security on the Internet. Using Tor can help you anonymize web browsing and publishing, instant messaging, IRC, SSH, and other applications that use the TCP protocol. Tor also provides a platform on which software developers can build new applications with built-in anonymity, safety, and privacy features.
The Tor project was launched by The Free Haven Project in 2002. In the past, Tor development was funded by contracts with the Naval Research Lab (inventor of onion routing) and the Electronic Frontier Foundation (who still kindly hosts our website). Sponsors of Tor get personal attention, better support, publicity (if they want it), and get to influence the direction of our research and development!


This is the ninth development snapshot for the 0.1.1.x series. We fix some memory leaks from the last snapshot, remove a lot of extra confusing log warnings, and fix other bugs.

http://tor.eff.org/download.html

AMD Dual-Core Linux Cluster deplyed at University of California.

well it is not at berkeley but at Davis. The University of California at Davis is an early InfiniPath customer and represents the very first AMD dual-core Opteron cluster deployed with PathScale InfiniPath.
Pathscale announcement was made at LinuxWorld in San Francisco.
The ultra-low latency and unprecedented messaging rate of the PathScale InfiniPath HTX Adapter greatly improves MPI application performance and Linux cluster utilization, enabling scientists, mathematicians and engineers to solve new problems with higher degrees of resolution than ever before. The highly pipelined, cut-thru design of InfiniPath is optimized for applications sensitive to communication latency, the most difficult problem to overcome when migrating from large SMP systems to clusters. InfiniPath delivers superior Interconnect performance at commodity price levels by implementing a high-performance software stack and connecting directly to the AMD Opteron™ processor via a standard HyperTransport HTX slot. When combined with low-latency InfiniBand switching from Cisco (TopSpin), Silverstorm (Infinicon) or Voltaire, InfiniPath enables applications to reliably scale to hundreds or thousands of nodes.
The Center for Computational Science and Engineering (CSE) at the University of California, Davis is concerned with the development of computational models and simulations as a means of understanding complex physical and biological processes, and to model and visualize entirely abstract processes encountered in physics, mathematics, engineering and computer science.
The Center for Computational Science and Engineering (CSE) at the University of California, Davis. was among the first to adopt the PathScale InfiniPath interconnect . CSE is implementing a 144-CPU AMD Opteron processor-based Linux cluster that leverages InfiniPath to run computational models and simulations related to physics, mathematics, engineering, biomedical diagnostics, and other processor-intensive HPC applications. This deployment consists of 36 server nodes from TeamHPC, a division of M&A Technology. Each server is equipped with two dual-core AMD Opteron processors and an InfiniPath HTX InfiniBand Adapter. They are interconnected with a Cisco TopSpin 270 InfiniBand switch.
The CSE also uses The PathScale EKOPath Compiler Suite at the facility.

Saturday, November 19, 2005

SuperComputing 2005 conclusion

I was going to write about the SC|05 and how it went for me. But then I happen to read one of my fellow blogger, Dan Ciruli's post at WestCostGrid, then realized that he had done it for me!
Too bad I did not meet him, over there at the SC|05, in fact I did not see much, I was doing too much work with servers and in the back rooms!
Take out the phrase like "my first trip" and "listened to Mr.Gates", He said most of what I wanted to say.
It was nice of him to write up about one of the institutions that I am affiliated! I let you guess what it is! I will write a specific article later on my technical experiences.

Friday, November 18, 2005

Grid Is not only for research and Computing, There is a log fun in Grids!

The G-cluster Ltd. was founded in 2000 to meet the demand for high-quality gaming on broadband network connected thin clients such as set-top boxes and next generation wireless devices. By moving most of the processing to sophisticated server clusters, G-cluster has been able to achieve network and device independence and exceptional cost-effectiveness without sacrificing the quality of graphics or gameplay. The G-cluster Game System is currently successfully being deployed in several commercial pilots and is ready for large-scale deployment. Here you will find some more information.
They have announced that an agreement with game industry giant Capcom Co. to bring Capcom's blockbuster franchises to Club iT's G-cluster Gaming-on-Demand platform.

The latest releases of Capcom's console hit series such as Resident Evil(R), Onimusha(TM) 3: Demon Siege, and MegaMan(R) will be added to the G-cluster gaming portfolio, along with other popular arcade titles. The games will be made available on both PC and digital set-top box (STB) services in Japan and other countries worldwide starting this December. The deal means that any user connected to the G-cluster service via a broadband network can easily enjoy these Capcom favorites.

First batch of games to be available on G-cluster includes;

Resident Evil(R) 2

Resident Evil(R) 3: Nemesis

Onimusha(TM) 3: Demon Siege

Breath of Fire(R) IV

Chaos Legion

Dino Crisis 2

Mega Man(R) X5

Mega Man(R) X8
I think that is a lot of fun!, If you play these games.

Wednesday, November 16, 2005

Grid Service Broker view Released by Gridbus Project


The New version 2.4 of the Gridbus Broker, supports for Globus toolkit version 4 and integration within portal environments. Also noted is that support for the use of multiple brokers within the same VM has been improved. Future releases of the broker will continue this trend towards increased support for portal and other eScience application developers needs.

At a Glance;
* Support for Globus 4.0 job submission services using WS_GRAM
* Improved stability for all middleware
* Improved persistence in order to support a wider range of databases.
* Introduced in-process and dynamic database creation via HSQLDB.
* New event-based Round-Robin scheduler that provides a simple scheduling algorithm
* Various bug fixes



It contains the broker classes, source code, and documentation. Please note that it allows you to schedule your compute and data-intensive applications on resources running Globus, Unicore, Alchemi , Condor, SGE, and XGrid.

If you are interested, you can become a member of the Gridbus community and wants to extend or make use of it, please let the team know. The broker will undergo continuous enhancement and the team welcomes your participation and encourages you to become a co-author of future versions of the software.
For information about requirements and installation please refer to the Gridbus broker manual(PDF).

The Team, who's hardwork brings GRID Technology to Masses;

* Rajkumar Buyya (Chief Investigator)
* Srikumar Venugopal (Lead researcher and developer)
* Krishna Nadiminti (Developer) (kna@csse.unimelb.edu.au)
* Hussein Gibbins (Developer) (hag@csse.unimelb.edu.au)
* TianChi Ma (Developer) (tcma@csse.unimelb.edu.au)
* Marcos Assun (Developer) (marcosd@csse.unimelb.edu.au)

VOIP IP Telephony: We all need to know about patents, Now we have a place to go

VOIP IP Telephony: We all need to know about patents, Now we have a place to go

Tuesday, November 15, 2005

Grid MP(TM) software suite has been certified on Hewlett Packard's Cluster Platform 3000 and Cluster Platform 4000 and Blade Server cluster

United Devices announced today that its Grid MP(TM) software suite has been certified on Hewlett Packard's Cluster Platform 3000 and Cluster Platform 4000 and Blade Server cluster at HP's Houston Partner laboratory complex.

Press Release BusinessWire
AUSTIN, Texas--(BUSINESS WIRE)--Nov. 14, 2005--United Devices announced today that its Grid MP(TM) software suite has been certified on Hewlett Packard's Cluster Platform 3000 and Cluster Platform 4000 and Blade Server cluster at HP's Houston Partner laboratory complex.

Testing at the HP laboratory complex was conducted using the Novell SLES 9 Validation Suite including the Novell(R) Validated Configuration Program, which defines and certifies multi-application, integrated application and platform stacks on Linux. Moving beyond current industry certification programs, the Novell Validated Configuration Program addresses an integrated stack when multiple components must work together. As a result, customers can smooth deployment of complex, mission critical Linux-based applications.

"Grid MP provides several competitive advantages to Linux cluster users," said Ben Rouse, United Devices' chief executive officer. "The Grid-Ready(TM) capabilities for cluster extension now let customers provision and share their HP Cluster Platform resource without purchasing additional grid management packages. And unlike other grid technologies, Grid MP can schedule work to clusters managed by multiple third party resource managers like LSF, PBS or SGE."

Grid MP(TM) from United Devices is the only infrastructure management solution that allows organizations to fully leverage and extend the processing power of Linux clusters. A Grid MP-enabled solution can power a cluster-only installation and then easily scale as organizations add both dedicated and non-dedicated resources -- including nodes within the cluster, other clusters, or any compute resource within the department.

This "grid-ready" feature of Grid MP allows other nodes to be added and managed easily without manual interventions or procuring additional software packages. In addition, UD's meta-scheduling capabilities enable workload management across a range of third party distributed resource managers (DRMs) already operating on existing clusters.

"The expanded capabilities announced with United Devices come at the right time for cluster users," said Bruce Toal, director of marketing at HP. "As demand for high-performance computing power continues to grow, cluster solutions need to be flexible enough to support a wider range of customer needs. By bumping the ProLiant Cluster Platform up to 64-bit and adding Novell SUSE support, we've teamed with United Devices to provide a powerful new resource for the cluster market."

The Novell Validated Configuration Program develops fully tested, fully integrated configurations that address enterprise business problems. Each of these configurations will provide Novell customers with all technical and business details necessary to model and accelerate Linux-based solution deployments.

Expanding Novell's recently announced data center strategy, the Novell Validated Configuration Program focuses on security, high availability, virtualization, systems management and monitoring, and application hosting. The program also supports high performance computing architectures for large-scale transaction processing environments.

About United Devices

United Devices is a market leader in grid and infrastructure management software and services. The company's products optimize an organization's existing resources (including clusters, servers and desktops) to create an enterprise-wide capacity on-demand environment. United Devices also operates the world's largest grid for grand-scale research that consists of over three million systems in more than 191 countries and territories.

United Devices, Grid MP and Grid-Ready are trademarks of United Devices.

Contacts


The Alliant Group for United Devices
Steve Eames, 210-698-2580
seames@thealliantgroup.com

Parallel Applications Software Development Kits (SDKs) for High Performance Computing Announced!


Supercomputing 2005
Absoft's New High Performance Computing Software Development Kits Open New World of Parallel Application Possibilities for Developers On 64-bit Intel(R) Xeon, and IBM POWER(TM) Systems


Press release Businesswire
SEATTLE--(BUSINESS WIRE)--Nov. 15, 2005--

Absoft's New High Performance Computing Software Development Kits Open New World of Parallel Application Possibilities for Developers On 64-bit Intel(R) Xeon, and IBM POWER(TM) Systems


Absoft Corporation, leading provider of software development tools, announced today at the SC05 Conference in Seattle new additions to its increasingly popular family of High Performance Software Development Kits. The Absoft High Performance Computing Software Development Kits (HPC SDK) are complete development solutions that allow software developers to easily compile, run, debug, and optimize high performance software applications on clusters, grids, and multiprocessor/multi-core systems built using 64-bit Intel(R) Xeon(R), and IBM POWER(TM) processors. The Absoft SDKs are the only HPC software development toolkits available today, and are offered exclusively by Absoft for systems running Linux, OS X, and soon, for Microsoft Windows Compute Cluster Solution(TM) (CCS).

Each HPC SDK includes the industry's leading software development tools optimized for 64-bit extensions and is preconfigured for easy installation. Every SDK includes Fortran and C/C++ compilers, debuggers, math libraries, pre-built message passing (MPI) libraries, and other development tools. Also included are a comprehensive set of example HPC programs to help new users quickly learn how to take full advantage of the HPC SDK features on their cluster, grid, multiprocessor, or multi-core system. Every HPC SDK comes with twelve months of continuing maintenance for all included components. The HPC SDK from Absoft is the only commercially available HPC SDK that combines tools from multiple vendors to provide customers with a true best-of-class solution.

"We have been using your SDK for just over 6 months now and we have never had a better suite of compilers and tools. The HPC SDK on POWER delivers ease of installation and usage, high performance, flexibility, and reliability that is unsurpassed on the Linux ppc64 platform. Using the HPC SDK Enhanced Edition with its innovative and comprehensive compiler, debugger and development technologies has allowed researchers at the University of Oregon's Neuroinformatics Center to improve their application performance and has streamlined our development process dramatically. We have used the SDK in neuroscience software development, including EEG signal processing and computational head modeling. In addition, we are applying the suite in our parallel performance tools research and development as part of the TAU Performance System project. In our experience, the HPC SDK Enhanced Edition for IBM Linux on POWER has required the lowest maintenance overhead combined with the best feature set for our ppc64 Linux development platforms. The compilers are robust, the parallel libraries are well integrated," said Dr. Allen Malony, Director of the Neuroinformatics Center at the University of Oregon.

Absoft announced the following new HPC SDK versions along with additional high performance tools:

-- HPC SDK Basic Edition 2.0 for 64-bit Intel(R) Xeon(R)

-- HPC SDK Enhanced Edition 1.0 for 64-bit Intel(R) Xeon(R)

-- Intel VTune(R) Performance Analyzer

-- Intel Thread Checker

-- HPC SDK Basic Edition 2.0 for IBM Power(TM)

-- HPC SDK Enhanced Edition 1.0 for IBM Power(TM)

-- IBM High Performance Computing Toolkit for IBM Power(TM)

-- IBM FDPR-Pro post-compilation optimization tool for feedback-directed program restructuring

"Absoft continues to expand our comprehensive, exclusive line of SDK solutions for high performance software application developers. As seen by our customers' testimonials, the Absoft HPC SDKs are innovative toolkits which provide excellent value," said Jeff Livesay, Absoft chief operating officer. "We are experiencing increasing traction in this exciting growth market and have many new powerful tools now shipping or soon to be available. Absoft is listening to HPC customers' expression of their needs and providing the complete, commercially-supported solution bundles to meet those needs."

HPC SDK Basic Edition 2.0 for 64-bit Intel(R) Xeon(R)

With the success of the first edition, Absoft is proud to announce the HPC SDK Basic Edition 2.0 for 64-bit Intel(R) Xeon(R) processor-based servers and clusters running Linux. This HPC SDK is updated with compilers and libraries from Intel(R) including C++ and Fortran 9.0 compilers, Intel(R) MPI, and Intel(R) Cluster Math Kernel Library 8.0. This HPC SDK also includes updated LAM/MPI and new Absoft examples and benchmarks.

The HPC SDK Basic Edition 2.0 for 64-bit Intel(R) Xeon(R) processor-based servers and clusters running Linux fully supports the recently announced IBM Grid and Grow(TM) hardware and services offering. This Grid computing support makes it easier for companies of all sizes to build and deploy affordable and efficient Grid computing IT operations. (http://www.absoft.com/corporate/pressreleases/051004_01.html)

"The IBM Grid and Grow solution based on the 64-bit Intel Xeon(R) processor provides end users a cost-effective foundation for a low-risk quick-start deployment of Grid technology. And, the close collaboration between IBM, Intel Corporation and Absoft resulted in a 64-bit Intel Xeon(R) processor-based HPC SDK, which makes it easy for companies, small and large to use the Grid and Grow foundation to build applications that are solid and high performing," said Robert Fogel, Director of Worldwide Grid Strategy and Business Development, Intel Corporation. "The combination of the Grid and Grow foundation plus the applications developed with the Absoft SDK delivers an innovative and scalable IT solution that can deliver high-impact business value."

HPC SDK Enhanced Edition 1.0 for 64-bit Intel(R) Xeon(R)

The HPC SDK Enhanced Edition 1.0 for 64-bit Intel(R) Xeon(R) processor-based servers and clusters running Linux includes all of the features of the popular HPC SDK for Intel(R) Xeon(R) systems with additional tools from Intel(R), including, VTune(TM) Performance Analyzer to easily streamline code, and Thread Checker for OpenMP debugging support. Also included are open source tools such as Marmot, an MPI diagnostic tool. Future versions of the Enhanced Edition of the HPC SDK will feature the FxP parallel MPI debugger from Absoft.

HPC SDK Basic Edition 2.0 for Linux on IBM Power(TM)

The updated HPC SDK Basic Edition 2.0 for Linux on IBM Power(TM) processor-based servers and clusters running Linux receives many additions, such as IBM's XL C/C++ V8.0, IBM's XL Fortran V10.1 compilers, and the Engineering Scientific Subroutine Library (ESSL). This HPC SDK also includes updated LAM/MPI and MPICH2, and new Absoft examples and benchmarks.

Introduced at the inaugural GridWorld(TM) Conference in Boston, this Absoft High-Performance Computing Software Development Kit (HPC SDK) has been specifically designed to work with the IBM Grid and Grow(TM) hardware and services offering. IBM Grid and Grow is a powerful solution designed to help organizations get started quickly with affordable Grid computing.

HPC SDK Enhanced Edition 1.0 for Linux on IBM Power(TM)

The HPC SDK Enhanced Edition 1.0 for Linux on IBM Power(TM) processor-based servers and clusters running Linux includes all of the features of the Basic Edition with the addition of powerful tools from IBM, including the IBM FDPR-Pro tool and the IBM High Performance Computing Toolkit. See the following sections for descriptions of these tools.

FDPR-Pro

The FDPR-Pro tool, developed by the IBM Haifa Research Lab, uses Feedback Directed Program Restructuring to achieve performance gains of anywhere from 5%, 10%, or more for software applications that have already been compiled and linked. FDPR-Pro for Linux on Power is a post-link utility for improving the performance applications that were compiled for the Linux operating system running on Power family platforms.

The tool receives input files in XCOFF format, instruments them, executes them for profiling information, and then optimizes them to achieve:

-- Better hit/miss i-cache ratio

-- Reduced number of branches

-- Reduced number of TLB misses

-- Reduced number of page-faults

The FDPR-Pro tool for Linux on IBM Power handles very large executables (such as an Oracle database application) and shared libraries. The tool is supported for both 32-bit and 64-bit Linux on Power for the following configurations:

Operating systems: Suse/SLES9 Redhat/EL4
Compilers: GCC 3.3(*), GCC 3.4(*), XL C/C++ 7.0,
XL Fortran 9.1
Processor architectures: Power4, Power5, PPC970, PPC440 (embedded)

The FDPR-Pro application optimization tool is included in the HPC SDK Enhanced Edition 1.0 for IBM Linux on Power(TM) and is also offered as a standalone product. FDPR-Pro will be generally available exclusively from Absoft in the first quarter of 2006. Developer licenses retail at $2400 for up to four processors, with limited time introductory discounts of up to 50%. Academic discounts are also available. Detailed product information and an application for limited introductory licenses and pricing discounts can be found at http://www.absoft.com/Products/Tools/fdpr-pro/

IBM High Performance Computing Toolkit

The IBM High Performance Computing Toolkit (HPCT), developed at the Advanced Computing Technology Center (ACTC), part of IBM's T.J. Watson Research Laboratory in Yorktown Heights, New York, is a suite of performance-related tools and libraries to assist in application tuning. This toolkit is an integrated environment for performance analysis of sequential and parallel applications using the MPI and OpenMP paradigms. It provides a common framework for a number of IBM systems offerings, including IBM eServer pSeries(R) systems. The IBM HPCT includes tools for:

-- Hardware Performance

-- Shared Memory Performance

-- Message-Passing Performance, and

-- Performance Visualization

"We use the IBM HPC Toolkit at NERSC, and find it to be invaluable in understanding the performance behavior of our applications on our IBM pSeries systems" according to Dr. David Skinner, HPC Senior Applications Specialist at NERSC. "We especially like the Toolkit's ability to trace back the performance data to the actual source code statements of our applications in the highly intuitive and visual framework."

Included in the HPC SDK Enhanced Edition 1.0 for Linux on IBM Power(TM), the IBM HPCT is also offered as a standalone product and will be generally available exclusively from Absoft in December, 2005. Introductory licenses are available starting at $5,000; site licenses and academic discounts are also available. Detailed product information and an application for limited introductory pricing discounts are available at http://www.absoft.com/Products/Tools/hpc-toolkit/.

For more information on Absoft's complete line of HPC SDK solutions, please visit: http://www.absoft.com/Products/Clusters/hpcsdk.html.

About Absoft Corporation

Celebrating 25 years of profitably in offering software solutions, Absoft Corporation (Rochester Hills, Michigan) is a recognized leader in software development tools for Fortran, C/C++, and high-performance computing on Microsoft Windows, Apple OS X and Linux on 32-bit and 64-bit processors, from AMD, IBM and Intel. Absoft is an active business partner with AMD Apple, IBM, Intel, Microsoft, Motorola, and other major technology companies, with active customers using Absoft products numbering in the tens of thousands world wide, including installations at all major national laboratories, research institutions, universities, and Fortune-100 companies. Absoft products are distributed globally through more than 200 resellers.

For more information, please contact Jeff Livesay at (248) 853-0050 or jal@absoft.com or visit www.absoft.com.

Trademarks

Absoft(R) is a trademark of Absoft, Corporation. All rights reserved. All other company/product names and service marks may be trademarks or registered trademarks of their respective companies. Absoft Corporation reserves the right to change the specifications, functions, or features, at any time, without notice.

Contacts


Absoft Corporation
Jeff Livesay, 248-853-0050
jal@absoft.com
www.absoft.com

Windows Compute Cluster Server 2003 get infiniband makeover

Voltaire Delivers Complete InfiniBand-Based Solution for Microsoft Windows Compute Cluster Server 2003


Supercomputing 2005

SEATTLE and BILLERICA, Mass.--(BUSINESS WIRE)--Nov. 15, 2005--

HLRS Selects Solution for High Performance Computing Applications

Press Release BusinessWire
Voltaire, the worldwide leader in grid backbone solutions, today announced a complete InfiniBand-based switching solution and software stack to support Microsoft Windows Compute Cluster Server 2003. The Voltaire solution enhances the scalability and simplifies the deployment of Windows-based clusters -- increasing communications of the systems 10 to 60 times over traditional methods. The solution is ideal for commercial high performance computing (HPC) applications employed by industries including scientific engineering, oil and gas exploration, automotive and financial services.

Windows Compute Cluster Server 2003 is Microsoft's initial offering for high-performance computing, and is designed to accelerate time-to-insight by providing an HPC platform that is simple to deploy, operate, and integrate with existing infrastructure and tools. Voltaire, an HPC leader, offers multi-service switches with integrated InfiniBand, Gigabit Ethernet and Fibre Channel connectivity that are deployed in many of the world's largest supercomputers and grids.

"Microsoft is pleased to be working with Voltaire to deliver a high-quality clustered computing experience to our customers," said Kyril Faenov, director of high performance computing at Microsoft Corp. "Voltaire's InfiniBand-based Grid Backbone switching solutions bring enhanced performance, scalability and manageability to Windows clustered solutions."

HLRS, a Microsoft Institute for High Performance Computing in Stuttgart, Germany, is currently using the Voltaire Grid Director ISR 9288 switching platform with Windows Compute Cluster Server 2003 on a 200-node cluster. HLRS provides support in the use of leading-edge supercomputer technology and applications to HPC users from the R&D organizations of leading companies in the automotive and other industries.

"Our goal is to simplify HPC tools for mass usage in a variety of industries," said Prof. Dr.-Ing. Michael M. Resch, HLRS. "The combination of Windows Compute Cluster Server 2003 and Voltaire Grid Backbone switching solutions offers both simplicity and high-performance in a single, elegant solution."

SC|05 Demonstrations

Voltaire and several of its system vendor partners will demonstrate the new solutions for Microsoft Windows Compute Cluster Server 2003 in the Microsoft booth (# 1206, 1208) at SC|05. Voltaire, HP and Ansys will show the Ansys Workbench and Ansys Distributed Solver using Microsoft MPI on Windows x64. In addition, there will also be a demonstration that will show a state-of-the-art LS-Dyna crash analysis simulation using Microsoft Windows Compute Cluster Server 2003, Voltaire, and NEC blade servers.

"We look forward to working with Microsoft to deliver high-performance solutions to address scientific and business computing challenges," said Patrick Guay, vice president, marketing, Voltaire. "Customers looking to run commercial HPC applications on Windows using clusters and grid-based configurations now can scale out and get great performance at a very attractive price point."

Voltaire solutions for Microsoft Windows Compute Cluster Server 2003 will be generally available in the first half of 2006 through Voltaire's OEM, integrator and reseller partners.

About Voltaire

Voltaire is the worldwide leader in grid backbone solutions for networked computing in the next generation data center. Voltaire's integrated family of switching hardware and network virtualization software delivers the high performance, intelligent backbone for grid computing architectures. Leveraging the InfiniBand standard, Voltaire solutions offer improved performance, utilization and scalability across compute clusters, storage and IP networks. Voltaire solutions are available from major systems vendors and integrators. More information about Voltaire is available at www.voltaire.com or by calling 1-800-865-8247.

All product and company names mentioned herein may be the trademarks of their respective owners.

Contacts


Christy Lynch
Voltaire
(617) 794-1362
christyl@voltaire.com

Sunday, November 13, 2005

HPC Analytics Challenge



"The benefit of the HPC Analytics Challenge contest is to provide a forum for researchers, engineers, and analysts to showcase data-intensive applications that solve real-world, complex problems through the use of rigorous and sophisticated methods of data analysis and high-end visualization," said Donald R. Jones, Ph.D., Pacific Northwest National Laboratory and co-chair of the HPC Analytics Challenge.

The HPC Analytics Challenge is an award competition honoring select technical and commercial applications that use leading-edge, advanced analytics techniques to solve complex, real-world problems. The growing need to leverage large amounts of underutilized data has led to the use of sophisticated methods for analysis and high-end visualization in conjunction with high performance computing, bandwidth and networking capabilities. The HPC Analytics Challenge is a unique opportunity for researchers, engineers, and analysts to showcase innovative techniques of rigorous data analysis (for example, forecasting, data mining, optimization, predictive analysis) and high-end visualization (for example, dynamic modeling, real-time rendering). Join prestigious analytics gurus from around the world to demonstrate your superior HPC analytics application. Learn more about this new and exciting competition by reading below.

Some of the key features important for applications to address in the HPC Analytics Challenge include:

* Innovative and sophisticated analysis and visualization techniques
* Performance and scalability (with large, complex data)
* Portability and distribute-ability
* Flexibility and deploy-ability
* Knowledge & discovery.

Additionally, the applications will be judged on the strength of the solution meeting the combination of various specific criteria such as:

* Utilizing various HPC technologies
o Analysis: Data mining, forecasting and optimization techniques, etc.
o Visualization: Interactivity of the application such as drill-down, real-time rendering and sophisticated visualization capabilities, etc.
o High Performance Computing
o Newer technologies such as web services, clusters, grids
o High Performance Storage
o High Speed Networking
* Implementing unique investigation and methods for data analysis
* Handling massive amounts of data
* Providing for collaboration, sharing, and ease of use of the application
* Achieving scalability and portability
* Providing significant and meaningful results from the use of the application

Applications for consideration ideally would focus on solving real-world, complex problems in a number of sectors. Some areas of:

* R&D/National Labs/ Homeland Defense, DOD, DOE (security, disaster planning)
* Oil and Gas (oil drilling optimization)
* Large Scale Science (astrophysics, High Energy physics, etc.)
* Finance (risk management, portfolio optimization)
* Corporate management (business analytics)
* Life Sciences or Healthcare (clinical trials, pharma, bio-medical research)
* Environment (predicting weather, forecasting, remediation)
* Web search engine enhancements
* SPAM and Intrusion detection or mal-ware detection

Participants are encouraged to provide their own dataset-"sanitized" as appropriate. Details of dataset will need to be provided to judges, including

* Data set size
* Complexity
* How did you use this dataset

The following links provided are readily available, public datasets for participant use if so desired. Alternatively, participants can use their own data sets.

* Bathymetry
geopubs.wr.usgs.gov/open-file/

* Census
www.census.gov/geo/www/tiger/tgrcd108/tgr108cd.html

* Geological and Climatological
earthobservatory.nasa.gov/
seamless.usgs.gov/
glcf.umiacs.umd.edu/index.shtml
www.ngdc.noaa.gov/seg/cdroms/grav99v1/document/html/tbase.htm
www.ngdc.noaa.gov/mgg/topo/globe.html
nedwww.ipac.caltech.edu/
www.ncdc.noaa.gov/ol/satellite/satelliteseye/cyclones/pfctstorm91/pfctstorm.html
www2.lib.udel.edu/subj/godc/internet/epubs/climd/
climate.ntsg.umt.edu/html/climate_data.html

* GISuser Data Directory
libraries.mit.edu/gis/data/index.html

* Lidar
www.csc.noaa.gov/crs/tcm/index.html

* Variety of Data (U.S. - based)
nationalatlas.gov/atlasftp.html

Winning entries will receive a special certificate and a nominal monetary prize and will present their work during the contest session at the conference. The length of the oral presentation is still to be determined but is likely to be similar to a short paper or presentation as part of the HPC Analytics Challenge session in the Technical Programs schedule. Also, plans and preparations will be undertaken such that winning entries would also show their applications "live" at the SC|05 Conference.

We will award prizes for best entries in several categories. The judges will decide on the exact categories based on the material received - possibilities include "Best Overall", "Best Student Entry", and "Most Original".

Honorable Mentions will be forwarded on to the Technical Programs Committee for consideration for being presented as interactive posters at the conference.

Important Note: All award winners are required to attend SC|05 and participate in the HPC Analytics Challenge activities.

HPC Analytics at Super Computing 2005

Since this is the SC|05 week, I will post some information from the SC|05. HPC Analytics is one of the most interesting section of the super computing, grids, distributed computing and even lowly clusters.

"I expect HPC Analytics to be one of the most exciting initiatives at SC|05", said Bill Kramer, SC|05 General Chair of the conference. "HPC Analytics adds to the significance of SC|05's HPC Computing, Network and Storage efforts, bringing these resources together while adding sophisticated analysis methods to solve some of societies' most challenging problems."

A new, special initiative at SC|05, HPC Analytics will highlight rigorous and sophisticated methods of data analysis and visualization used in high performance computing by showcasing powerful analytics applications solving complex, real-world problems.

Upgrades in systems, bandwidth and networking technologies over the last decade have resulted in dramatic increases in performance, scalability, and overall computational power in high performance computing. Additionally, more than ever before, organizations in commercial, government, university, and research sectors are increasingly tasked with making sense of huge amounts of underutilized data. These dynamics have led to the growing area of HPC analytics. In addition to being a top priority in supercomputing communities worldwide, HPC Analytics is also a major influence on the growing trends of "Business Analytics" which includes rigorous data mining techniques and more predictive, forecasting analytics capabilities.

The primary goals of the HPC Analytics Initiative at SC|05 are the following:

* Sponsorship of the HPC Analytics Application Challenge to showcase select, powerful analytics applications solving complex, real-world problems
* Encouragement of Technical Program submissions highlighting HPC Analytics approaches including sophisticated data analysis and visualization

The HPC Analytics Initiative is supported by a group of volunteers from government, industry, and academia. If you are interested in joining or have any questions,

Saturday, November 12, 2005

Supercomputing 2005 is in Session


If you are into Grid Technology Supercomputing and or any of the HPC (High performance Computing), then this is the place to be on this weekend.
The event will continue until 16th if this month. There are exhibits, team challenges and Video conferencing at HPC level (See an article published on VOIP blog SNAPVOIP. VOIP IP Telephony: Video conferencing in Supercomputing 2005) There are much more than I can include here. So make sure you visit the site (Physical or Virtual!)
SC|05, the premier international conference on high performance computing, networking and storage, will convene Under the theme, "Gateway to Discovery," and will showcase how high performance computing, networking, storage and analysis lead to advances in research, education and commerce. ACM and IEEE Computer Society are the SC|05 conference series sponsors.

Since first SC conference in 1988, exciting technical and education programs, workshops, tutorials, an expanded exhibit area, demonstrations and many other activities have been added. SC|05 is the one place that attendees can see tomorrow's technology being used to solve world-class challenge problems today.

The last I attended was in 2003!
Phoenix, AZ November 15-21, 2003


Schedule for this week at SC05 is here.
Conference committee email addresses are arranged alphabetically here. Please send correspondence to the appropriate contact for more information on SC programs and initiatives.

Thursday, November 10, 2005

Oracle(R) Grid Index rises 5.2 from 4.4 Since April 2005

REDWOOD SHORES, Calif., Nov. 7 /PRNewswire-FirstCall/ --


Oracle (NASDAQ:ORCL) today unveiled the results of the latest Oracle(R) Grid Index research, which is dedicated to mapping the world's journey to Grid Computing. At a global level, the Oracle Grid Index has risen to 5.2 from its previous value, in April 2005, of 4.4. All of the underlying indices making up the overall Grid Index -- Foundation Readiness, Knowledge & Interest and Adoption Lifecycle -- show increases in all geographies, indicating steady progress of mainstream activity towards the implementation of modern dynamic Information Technology (IT) infrastructures.
(Logo: http://www.newscom.com/cgi-bin/prnh/20020718/ORCLLOGO )

As well as chronicling the Grid Computing journey, the research(1) illuminates strong links between business management and IT strategy. A key finding is that board-level involvement in IT strategy goes hand-in-hand with business success -- 72 percent of organizations in which business executives are routinely involved in IT strategy are growing, compared to just half that number in which executives are never involved. Furthermore, there is strong evidence that executive involvement in IT strategy and planning leads to better alignment of objectives, priorities and activities which in turn leads to a higher proportion of IT budgets being spent on the creation of business value rather than simply housekeeping.

Global Trends in Grid Computing

Examining the changes in Grid Index values from the previous instance (April 2005) reveal several notable facts:

-- USA makes most progress. With a jump in Oracle Grid Index from 4.6 to
6.1, the USA shows the biggest rise in Index values, which brings it to
join the Nordic countries as featuring the highest Index number. North
America also shows the greatest jump in Lifecycle Adoption sub-indices,
which indicates that actual adoption of Grid Computing is progressing
faster there.
-- Top Three remain unchanged. The Nordics, South-East Asia and the USA
continue to lead the rest of the world on the Grid journey, with Oracle
Grid Indices of 6.1, 5.9 and 6.1 respectively.
-- Southern Europe and traditional Asian markets move slowest. While
India has made a big jump, from 2.9 to 4.4, the other countries which
showed lowest Grid Index figures in April 2005 have moved most slowly
-- Greater China from 4.7 to 4.9, Italy from 4.0 to 4.2, Korea from 4.4
to 4.7 and Spain from 4.2 to 4.5. Japan, included in this survey for
the first time, makes its entrance at an Oracle Grid Index of 4.4.
-- Early Adopters Go For Enterprise Grid Solutions. Informed IT
professionals see Grid as being most relevant as part of their
corporate IT infrastructure. Among the respondents, an "experienced
guru" group(2) highlights cluster grids as being relevant to their
organization in 44% of cases, and a possible option in a further 52%.
Strength of feeling is even higher for distributed enterprise grids,
where the equivalent numbers are 69% and 29%. Far fewer see the firm
relevance of managed hosted grids (22%), utility grid services (25%)
and partner/community grids (7%).


"All indicators in all regions show positive progress in the awareness, understanding and adoption of Grid Computing," said Robert Shimp, Oracle's vice president of Technology Marketing. "The research proves that increasing numbers of organizations around the globe are taking notice of this market- changing technology. It also underlines that more and more of them are taking action to derive benefit from it in the form of better usage of current IT assets and better positioning for the future."

The Business Value of IT

The research also examines the attitudes of organizations' business management towards IT. As well as a clear correlation between business growth and regular business management involvement in IT strategy, the research reveals that IT is increasingly being considered as a core part of business investment:

-- Only 30 Percent of Organizations Can Rely on Business Executive Input
Into IT Strategy. Senior business managers tend to get involved with IT
on an event driven basis, typically in reaction to something that has
occurred in the business world that needs IT support, or when an IT
related event such as the systems failure has a negative impact on the
business. In the majority of organizations, IT appears to be thought of
more as a service function than as an integral part of business
operations.
-- High Business Involvement in IT Strategy is Linked to Better Use of IT
Investment. Traditional wisdom has it that the vast majority of IT
departments' workloads involve housekeeping and maintenance projects.
However, in over 60 percent of those organizations where business
management is always involved with IT strategy, over half of the IT
expenditure relates to business projects. The other end of the
spectrum does reflect the traditional wisdom, though; where there is no
business involvement in IT strategy, nearly 90 percent of
organizations use less than half of their IT spend on business
projects.
-- Business Executives Who Are Fully Involved in IT Strategy Appreciate
the Need for Sound Infrastructure. In the 31 percent of organizations
where IT departments can rely on input from business executives into IT
strategy, over three quarters of the executives concerned fully
appreciate the need for a sound IT infrastructure. At the other
extreme, less than 15 percent of those business executives that have
no involvement in IT strategy have a clear understanding of the role IT
infrastructure plays.


"The Oracle Grid Index research has some important messages for business people everywhere," Shimp continued. "The strong link that it shows between businesses which are growing and those whose executive management are always involved in IT strategy suggests that the board should regard IT as a strategic investment, not merely as a cost center. "

Strong Growth Continues in Key Underlying Technologies, Led by Asia Pacific

Comparison with previous Oracle Grid Index research shows continued strong growth in significant underlying technologies such as Service Oriented Architectures (SOA) and Blade Servers. Asia Pacific, whilst still lagging the other two regions surveyed in actual adoption numbers, has shown higher growth in both areas:

-- Service Oriented Architectures. The number of organizations currently
using or considering the adoption of SOA have risen 18.4 percent in
North America, 17 percent in Europe and 43.6 percent in Asia
Pacific(3).
-- Blade Servers. Organizations currently using or considering adoption
of Blade Servers rose dramatically. The number of such North American
organizations grew 91 percent in North America, 68.7 percent in Europe
and no less than 207.9% in Asia Pacific(4).

About Oracle

Oracle is the world's largest enterprise software company. For more information about Oracle, visit our Web site at http://www.oracle.com/.

Trademarks

Oracle, J.D. Edwards, and PeopleSoft are registered trademarks of Oracle Corporation and/or its affiliates. Other names may be trademarks of their respective owners.

The above is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decision. The development, release, and timing of any features or functionality described for Oracle's products remains at the sole discretion of Oracle.

(1) See "Notes To Editors" for details of the research

(2) 105 respondents out of the total of 1,466 declared the highest level
of grid computing knowledge (five on scale of one to five) along with
current experience of grid technology.

(3) SOA -- Proportion of North American organisations using or considering
SOA rose from 55.4 percent to 65.6 percent; European from 49.5 percent
to 57.9 percent; and Asia Pacific from 26.6 percent to 38.2 percent.

(4) Blades -- Proportion of North American organisations using or
considering Blade Servers rose from 34.7 percent to 66.6 percent;
European from 33.2 percent to 56 percent; and Asia Pacific from 14
percent to 43.1 percent.

Notes to Editors

Quocirca conducted 1466 interviews with senior IT influencers and decision makers, completed in September 2005. 302 of these were conducted in North America, 603 in Europe and the remainder (561) in Asia Pacific. Respondents were from a mixture of large multinationals and medium to large national organizations across a broad cross section of industry sectors.

Another way to fight Patents!

I was browsing at Groklaw and this caught my eye. "Open Invention Network Begins Stockpiling Patents - UPDATED" and I thought another greedy bunch had started to take part in money hunt. But as I read the article, I was glad and made me smile, Thank you PJ.
The new nonprofit company, called Open Invention Network, is receiving an undisclosed amount of money from International Business Machines Corp., Sony Corp., Philips Electronics NV, Novell Inc. and Red Hat Inc. It plans to buy Linux-related patents, offering royalty-free licenses to companies and individuals that pledge not to assert their own patents against the network's other licensees. . . .
Well Yahoo News has this article and New York Times has this to say.
After reading all the articles, I wonder now if nasty people will start patenting Linux stuff so they can sell it this company!
AAAAGH! It never ends, Patents are like Vitamin A, take a little, and you are fine. Eat a Dog's liver (High concentration of Vitamin A), you are dead

Wednesday, November 09, 2005

Cross-Continental InfiniBand Cluster Staged for Deployment at Supercomputing 2005


Supercomputing 2005

SEATTLE--(BUSINESS WIRE)--Nov. 9, 2005--The OpenIB Alliance today announced that nearly thirty organizations are confirmed to showcase the world's largest, cross-continental InfiniBand data center in conjunction with SCinet at next week's Supercomputing 2005 conference (SC/05) in Seattle, Washington (Washington State Convention and Trade Center). The SCinet InfiniBand cluster will host over 6TFlops of supercomputing performance, which would rank the cluster as high as 50 on the TOP500 list of the world's most powerful supercomputers. In addition, the SCinet InfiniBand fabric will have direct access to native attached InfiniBand storage solutions hosted by the StorCloud initiative. The strong industry mix of InfiniBand component vendors, infrastructure equipment suppliers, software vendors, and research and university end-users that are participating include:

AMD, Ames Laboratory, Appro, ASUS, Cisco, Cornell Theory Center, Dell, Emcore, Hewlett Packard, IBM, Intel, IWILL USA Corp., Lawrence Berkeley National Laboratory, Mellanox, Microsoft, NCSA, Lawrence Livermore National Laboratory, Los Alamos National Laboratory, Obsidian Research, Pathscale, Pittsburgh Supercomputing Center, Rackable, RedHat, Sandia National Laboratories, Silicon Graphics, SilverStorm, Sun Microsystems, Tyan Computer, and Voltaire

SC/05 participants with the OpenIB Alliance will use the OpenIB software stack for both Linux and Windows host systems, and the OpenIB subnet manager (OpenSM) will manage the entire Infiniband network. The availability of the OpenIB stack from all vendors is the industry's solution to widespread interoperability and mass deployment of InfiniBand fabrics in data centers for enterprise and performance computing.

The SCinet InfiniBand fabric locally interconnects participating booths on the SC/05 exhibit floor via InfiniBand over MPO multi-mode fiber using Emcore SmartLink modules.

The three remote InfiniBand clusters are at Lawrence Livermore National Laboratory in California, the Intel Dupont facility in Washington, and the Naval Research Laboratory in Virginia. These locations are connected through the wide area network (WAN) point of presence for SCinet, the Pacific Northwest Gigapop at Washington University, and utilize optical long-haul equipment from Ciena, Cisco, and Juniper running over networks provided by Abilene, Esnet, Internet2, National LambdaRail, and Qwest. At each endpoint, InfiniBand over optical (either DWDM or SONET OC192) is converted by Longbow XRs from Obsidian Research. The Longbow enables globally distributed InfiniBand fabrics to seamlessly cross connect by encapsulating 4X InfiniBand over OC-192c SONET, ATM, or 10GbE WANs at full InfiniBand data rates. The conversion is totally transparent to the InfiniBand fabric and is interoperable with OpenIB's software stack and subnet manager.

SC|05, the premier international conference on high performance computing, networking and storage, convene in November 2005 in Seattle. Under the theme, "Gateway to Discovery," SC|05 will showcase how high performance computing, networking, storage and analysis lead to advances in research, education and commerce. ACM and IEEE Computer Society are the SC|05 conference series sponsors.

Digipede Network 1.2 Released! (.NET Grid)

Digipede Technologies, introduced the Digipede Network Version 1.2, featuring full support for .NET 2.0, Microsoft's next-generation framework. There are two versions of the software, Team Edition and the Pro Edition. You can read on their site what version you need! There is also a SDK, providing developers with the tools and information needed to build "Digipede-enabled" master applications and distributed applications.. Digipede says its technology pools the power of Windows desktops and servers "to deliver dramatically improved performance for real-world business applications."

The solution is built entirely on Microsoft .NET technologies, and is easier to buy, install, learn, and use than competing grid solutions, according to the company.

Tuesday, November 08, 2005

Don't buy Sony!!

geemodo: No Sony Gadgets, Gizmo's or anything related to Sony on this site!

Fuji released Not the mountain, ROCKS 4.1

ASCRIBE News Letter:
SAN DIEGO, Nov. 2 (AScribe Newswire) -- Researchers at the San Diego Supercomputer Center (SDSC) and their collaborators have released version 4.1, code-named "Fuji", of the Rocks Clustering Distribution for immediate download (www.rocksclusters.org). Rocks supports i386 (Pentium/Athlon), single and dual-core x86_64 (Opteron/EM64T) and ia64 (Itanium) CPUs.

The major new feature in this release is the addition of the Avalanche Installer. Avalanche integrates peer-to-peer package serving mechanisms into the de facto standard Kickstart system to dramatically improve installation scalability. "The team has taken several tasks traditionally assigned to the front end and distributed them among the cluster nodes. With the addition of load-moderating mechanisms and the distributed package cache, we've tested 'fire-and-forget' re-installations on 100s of nodes from a single front end and now consider this scale routine," said Philip Papadopoulos, Program Director for Grids and Clusters at SDSC. In previous releases of Rocks, the front end dynamically created a complete Red Hat Kickstart file (the description for a node installation) and served all packages to each and every node. With the Avalanche Installer, the front end now delgates the majority of the dynamic configuration generation to the cluster nodes, while the BitTorrent-based peer-to-peer mechanisms take direct advantage of switched cluster network capacity for near-perfect scaling. Avalanche is only active while nodes are installing and has no impact on running nodes.

Since 2003, Rocks has supplied feature-specific Rolls to enable users to reliably and correctly customize their clusters. In Rocks 4.1, the default operating environment is encapsulated in the "OS Roll" which is based on CentOS 4 update 2 (a freely-available recompilation of Red Hat Enterprise Linux 4 update 2). The OS Roll can be substituted by any "true" Red Hat Enterprise Linux 4 (RHEL) distribution. CentOS 4, RHEL 4, and Scientific Linux 4 are all supported base operating systems for Rocks 4.1.

This release also includes refreshed packages for the Grid, Grid Engine, HPC, and Visualization Rolls. Grid-enabled endpoints are provisioned with The Grid Roll which now contains Globus Toolkit version 4.0.1. The Grid Engine Roll, which automatically installs and configures Sun's open-source Grid Engine job scheduling tool, has been updated to version 6.0u6. The HPC Roll, which contains essential tools for building computational clusters, now contains version 1.2.7 of MPICH for ethernet. Finally, the Visualization Roll, used to easily build tiled-display walls, has been refreshed with the latest SDL packages. In addition, several interconnect vendors have developed their own Rolls for Rocks. Rolls are available for Myrinet, Quadrics and Infiniband (SilverStorm Technologies, Voltaire and TopSpin).

The Rocks development community includes the Cluster Development Group at SDSC, Scalable Systems in Singapore, the HPC Group at University of Tromso in Norway, the SCE Group at Kasetsart University in Thailand, and the cluster development group at KISTI in Korea.

- - - -

CONTACT: Mason Katz, Rocks Manager, 858-822-3651, mjkatsdscdotedu

ABOUT ROCKS: Development of Rocks is funded from National Science Foundation, and aided by generous equipment donations from Sun Microsystems, Dell, AMD, SilverStorm Technologies and Intel. The Rocks project was started at SDSC in early 2000, with the goal of "making clusters easy". Today, the Rocks user base includes nine Top500 computers, and several hundred clusters around the globe. The Rocks Register, a web page where Rocks users voluntarily register their deployed cluster, shows that Rocks powers over 550 clusters with an aggregate of nearly 160 TFlops of peak computing. More information on Rocks, including documentation and complete access to the source code, can be found on the project's homepage (www.rocksclusters.org).

Media Contact: Mason Katz, Rocks Manager, 858-822-3651, mjkatsdscdotedu

Monday, November 07, 2005

Markit deploys an Oracle Database 10g and Real Application Clusters on a Linux server.

Markit originally evaluated the Oracle, Sybase, IBM and Microsoft databases before choosing Oracle Database 10g. It evaluated Oracle 10g RAC on a Linux platform in the scale lab. The result was performance improvements of between four and ten times compared with the original Sun Solaris combination.
The system will be used to manage over a million and a half market price checks daily for its 6,000 online customers.

Linux client is available to World Community Grid

World Community Grid of which I am a member of (Accumulated Points: 132,859 Current Ranking: 7,406) has finally released a Linux Client, via BOINC. A forum news release is here. and an excerpt is below.
The Linux client is available to World Community Grid members via the Berkeley Open Infrastructure for Network Computing (BOINC) platform. After much research, we found BOINC to be to best and most economically feasible solution for World Community Grid. For more information about BOINC, please see the Linux section of Help on our website.
To participate in the Linux program, go to the Join Now. From there, you will be directed to select either the Windows client or the Linux client. Once you select the Linux client, simply follow the instructions on your screen.

If you have any questions about Linux, there are four ways to get answers:
1. Review the Linux section of Help.
2. Go to the BOINC/Linux forum and review the Linux questions that others have posted.
3. Go to the BOINC/Linux forum and enter you own question and wait for a response.
4. Send an email to the support desk via the contact us feature.

Sunday, November 06, 2005

ObjectWeb takes Grid to China

Not really but it is collaboration between ObjectWeb and Orientware, an understanding to promote the adoption of open-source middleware worldwide.
ObjectWeb and Orientware announced their decision to build a common open source middleware platform and Orientware agrees to release several of its middleware components in open source: PKUAS, StarCCM, XLinker, WSWF and OnceTX. This cooperation between the two consortia is strongly supported by the Chinese Ministry Of Science and Technology and the French Ministry of Economy, Finance and Industry.
Press release is here.
Objectweb which I frequent is here.
Orientware, If you can read Chinese or want to use google translation, click

Friday, November 04, 2005

Make a bunch of clones, of your Linux machine/s!

I came across this article by Falko Timme while at SystemImager site. I have been using SystemImager but Falko has given a fine step by step guide. I think it is a good read if you are interested in replicating a few machines.

geemodo: What is GEEMODO?

geemodo: What is GEEMODO?

BLOGGERS Please Follow

This is from one of my other blogs. If you are here you should follow and read.
It is about printing press' attack on bloggers.

Thursday, November 03, 2005

F5 BIG-IP Load Balancer Beta Plug-in for Grid Control

This plug-in found here allows you to monitor F5 BIG-IP v9 Load Balancers to ensure Oracle stack performance.
This extension was developed by Oracle Corporation, and is not an F5 product.
Plug-in support for F5 BIG-IP v4.5 Load Balancers is already integrated into Release 2 of Enterprise Manager 10g Grid Control.
Press release here.

ODF served on a GRID! SUN is shining, Sit down and have a byte!

Using the new Sun Grid service, virtually any consumer with a Web browser will be able to upload proprietary documents, and have them automatically converted to Open Document Format (ODF). (The ODF is an XML-based industry standard file format specification for office productivity applications, including text, spreadsheet and graphical documents.)
Sun Microsystems intends to deliver on-demand network services to convert documents from native proprietary formats such as Microsoft Office into the Open Document Format (ODF), and also plans to deliver a service to convert text files to podcasts or audio files for playback at a later date. These introductions add more services to the company's catalog of grid services available via Sun's Grid Utility.
Sun plans to work with the open source community to enhance the OpenOffice and StarOffice platforms to leverage these services as native features. With more than 50 million downloads around the world, OpenOffice and StarOffice are the most rapidly growing productivity suites on the Internet, according to Sun.

Tuesday, November 01, 2005

Oracle Enterprise Manager 10g Grid Control (V2) released

Oracle Announces Oracle(r) Enterprise Manager Partner Offering - Enabling Broader Management of Systems and Services in Grid Environments
Press Release

Support for Leading IT Infrastructure Products
In Release 2, Oracle Grid Control boasts strong support for products from a number of leading technology vendors. With improved insight into the performance of partner products, Oracle Grid Control allows administrators to more effectively detect and resolve issues in a grid computing environment. Available today are new plug-ins that enable Oracle Grid Control to deliver diagnostic and management capabilities for:

* BEA Systems WebLogic (version 7.0 and above);
* F5 Networks BIG-IP Server Load Balancer (version 9.0 and above);
* IBM WebSphere (version 5.0 and above); and,
* NetApp(r) fabric-attached storage systems.

Plug-ins for Check Point Software's VPN-1(r) Pro with integrated firewall functionality and Juniper Networks Firewalls, among others, are planned.

"As organizations continue to move toward enterprise grids, a centralized management system becomes even more critical," said Paul Weinstein, vice president of business development for Check Point Software Technologies Ltd. (NASDAQ: CHKP). "Our integrated firewall and Virtual Private Network gateways coupled with Oracle Grid Control will enable us to offer our joint customers a secure, easily managed grid environment."

"Through this partner offering and the release of the BIG-IP Plug-in for Grid Control utilizing our iControl API, we are able to address customer demand for an integrated tool that monitors and configures the complete application and network infrastructure from one central location," said Jim Ritchings, vice president of Business Development at F5 Networks. "With Oracle Grid Control and F5's BIG-IP system customers now have access to highly available end-to-end application delivery networking, substantially lowering TCO."

"Enterprises are asking for our help in making systems manageability easier and less expensive," said Tony Scarfo, vice president of Global Alliances at Juniper Networks. "The Oracle Grid Control management plug-in for Juniper's Firewalls will provide our customers with a comprehensive tool for monitoring their entire IT environment."

"The combination of Oracle Grid Control and NetApp(r) storage systems provides our customers with a unique and systematic way to monitor their entire grid environment," said Patrick Rogers, vice president of Products and Partners at Network Appliance. "We are delighted to be the first enterprise storage platform to be managed and monitored directly from Oracle Grid Control. Now customers can monitor the performance of their storage processor utilization and disk I/O and review and adjust NetApp storage system utilization. This is another great example of how NetApp and Oracle are working together to make enterprise grids simple to manage and cost effective."

Enterprise Manager 10g Grid Control Release 2 provides a single tool that can monitor and manage not only every Oracle software element in your grid, but also Web applications, hosts, and the network in between. Grid Control is also extensible via an SDK so customers can use it to monitor additional components that are not supported out-of-the box. It is available for download here.