2007's first Top500 list is out leaded by BlueGene!

As I disclosed in my one later post dated Sunday, May 20, 2007 I am posting a yet another post on fastest supercomputer of the world. The list was published during at the International Supercomputing Conference (ISC ’07) in Dresden, Germany (http://www.supercomp.de) on June 27, 2007. Who else it can be other than BlueGene??

Yes, IBM BlueGene (http://top500.org/system/7747) installed at Lawrance Livermore National Laboratories, US (http://www.llnl.gov) topped on the list for the consequent forth time since its commissioning by delivering a stunning and steady 280.6 Teraflops! (Trillions of computations per second). In addition there were two other Supers which crossed the 100 Tflop mark -- the upgraded Cray XT4/XT3 at Oak Ridge National Laboratory, ranked No. 2 with a benchmark performance of 101.7 TFlops and Sandia National Laboratory’s Cray Red Storm system, which ranked third at 101.4 TFlops.

The unending π game

3.1415926535897932384626433832795028841971693993751058209749445923078164062862089986280348253421170679

This unending figure has always attracted philosphers, mathematicians, scientists, engineers and programmers ever since last 4 Millenia. A few intersting links of the efforts made by them are listed as below. Any addenda to the same are always welcome.

1. A very good history of these efforts : http://www.codehappy.net/pi.htm
2. An illustrative collection of different methods for calculating this magic figure: http://documents.wolfram.com/v5/Demos/Notebooks/CalculatingPi.html
3. Algortihms and code to achieve π (PI) for mathamatics and computer science experts: http://www.cygnus-software.com/misc/pidigits.htm
4. An incradible PI search page to find digits upto 200 million PI digits: http://www.angio.net/pi/piquery
5. Leibniz's formula on wikipedia : http://en.wikipedia.org/wiki/Leibniz_formula_for_pi
6. Collection of PI formula applied by various people: http://www.colab.sfu.ca/PiDay/3_14/PiTOC.html
7. A nice collection of binaries as well as sources programs for PI: http://myownlittleworld.com/miscellaneous/computers/pilargetable.html

Fedora 7 is released excluding the "Core"!!

After almost 24 months of continuous development and 6 concurrent core releases, Red Hat supported the Fedora was finally released as Fedora 7 alone excluding the core factor and combining the core and extras on May 31, 2007 (Just Yesterday!). It is considered to be the most stable version amongst its previous variants packed with varios features including various Spins, increased start-up and shut-down speed, secured wireless, support for encrypted FS, beautiful themes and many more to be found at http://fedoraproject.org/wiki/Releases/7/Features%3A?action=show.

In total F7 must be provided a few GB of space in your hard drive! (Now even live CD and a small bootable iso image is also available). So grab it now from here: http://fedoraproject.org/get-fedora.html

From Giga Flop to Tera Flop!

The performance delivered by any Super Computer has increased anything, like a rocket ever since their inception! Continuous implementations in microprocessor and semiconductor technology, interconnections and relatively the never failing Moore's law have always hiked the pick performance of a Supercomputer year by year.

Within just about a decade, now some of the top Supercomputers deliver their performance in Teraflops and hence, talking and measuring them in Gigaflops has become like a history. Top500.org (http://www.top500.org) is an official benchmarking group that measures the top notch delivery of 500 most powerful Supercomputers by using the high performance linpack (www.netlib.org/linpack) benchmark as a yardstick twice a year.

Let us now examine the peak performance of some the top500 #1 ranking Supercomputers in the last decade. :

This was the first phase:

• Intel Paragon XP/S @ Touchstone Delta System CalTech, 1993, 143.4 Gigaflops
• Fujitsu Numerical Wind tunnel @ National Aerospace Laboratory Japan, 1993, 170 Gigaflops
• Thinking Machine Corporation, CM-5, 1995
• Hitachi CP-PACS, 368.2 Gigaflops, 1996

Here comes the second phase: (Observe the hike in performance)

• IBM ASCI-Red @ Sandia National Laboratories, New Mexico, 1338 Gigaflops, 1997
• IBM ASCI-Red @ Sandia National Laboratories, New Mexico, 2121 Gigaflops, 1999
• IBM ASCI-White @ LLNL, 7304 Gigaflops, 2000

Here comes the third but slightly short phase:

• NEC, The Earth Simulator @ the Earth Simulator Center (ESC), Japan, 35860 Gigaflops, 2002

Finally Supercomputers delivers @ Teraflops, it is really increadible.

• IBM Blue Gene @ LLNL, 280.6 Teraflops (An increadible 280600 Gigaflops), on the top since 2004

Friends, this progress is just a glimpse of last 13-15 years, but the entire hardware as well as software technology behind this phenomena is now around almost 65 years old. To conclude this entery we have to remember Mr. Gordon Moore once again to just predict the future of Supercomputing!!

By the way the next list of top500 is to be launched in the June 2007 at ISC 07, Germeny (http://www.supercomp.de/). We will be back here again on this topic as the new list is declared..... till then wait for YABP!

Emulators Watch - 2007

Hey friends, we are going to discuss emulators useful for geeks and not for gamers (though many times a geek is found a gamer and vice versa!). Some of the links of popular emulators as follow.

• Bochs - http://bochs.sourceforge.net/
• Xen - http://sourceforge.net/projects/xen/
• Qemu - http://fabrice.bellard.free.fr/qemu/
• Wine - http://www.winehq.org/
• VmWare emulation products - http://www.vmware.com/
• Microsoft VirtualPC http://www.microsoft.com/windows/products/winfamily/virtualpc/default.mspx)

Free Operating Systems/Linux @ Education

This post in our blog may seem to be slight different from the previous ones. As it concerns with more to social issue rather than the technical ones. Although the Free and Open Source Software has always direct concerns with the society, more specifically liberalised and free society.

Well, there has been many educational Linux distributions which are specially engineered for the school going children. This distros are packed with educational games as well as easy to use scientific, mathematical and astronomical software which prove very much useful in their education. Some the links are as follow:

• Edubuntu, an Education linux by Ubuntu (http://www.edubutnu.org)
• MINIX III, (http://www.minix3.org)
• Freeduc, (http://www.ofset.org/freeduc-cd)
• Lormalinux by Lorma college, (http://linux.lorma.edu/)
• Linux for Kids, (http://www.filegate.net/linux4kids/)
• K12 Linux Terminal Server Project, (http://k12ltsp.org/contents.html)
• Quanitan Sceintific Computing Environment (http://dirk.eddelbuettel.com/quantian.html)
• Athena by MIT ,(http://web.mit.edu/ist/topics/linux/index.html)
• Seul, (http://www.seul.org/)
  

So this list is not ending here, there is lot contributed for education.

How SSI Cluster differ from PVM Cluster?

SSI:

SSI stands for Single System Image which is the simplest way of High Performance Computing, it uses a clustering middleware such as openMosix (http://openmosix.sourceforge.net/). In the SSI cluster, all the member nodes in a Local Area Network collectively deliver performance as single virtual machine.

Advantage:

• Very easy to setup.
• No special hardware and software requirements.
• Applications need not to be recompiled for multicomputing environment.
  

Usage:

Batch processing aaplications such as Backup(tar+zip) for mulitple files, Media Format Conversion for mulitiple files, Content filtering and e-mail virus scanning etc.

PVM:

PVM stands for Parallel Virtual Machine, in which the cluster distributes a single task into crunches and processes them individually over the member nodes of the cluster. However, a major issue with this type of custering is the recompilation of the application with the PVM/MPI (Parallel Virtual Machine/Massge Passing Interface) libraries such as OpenMPI (http://www.open-mpi.org/) or OSCAR (http://oscar.openclustergroup.org/).

Advantage:

• Can distribute a single task requireing higher CPU Throughput into multiple sub-tasks.
• Node job assignment and thread management is taken care by parallely compiled application itself.

Usage:

Applications that require more CPU crunching such as scientific applications, data modelling, weather forecasting etc.