Chinese supercomputer is world’s fastest at 2.5 petaflops

China set to claim supercomputing crown - October 28, 2010

In a potential blow to US national pride the world’s fastest supercomputer is now Chinese, beating the Americans into second place for the first time since 2004 with a machine which is smaller and more energy efficient than its closest US rival.

In the run up to the release of the official list of the top 500 supercomputers next week the Chinese supercomputer, Tianhe-1A, looks certain to occupy the top spot

...

The system uses 7,168 NVIDIA Tesla M2050 massively parallel graphics processing units (GPUs) and 14,336 multi-core central processing units (CPUs). It would require more than 50,000 CPUs and twice as much floor space to deliver the same performance using CPUs alone. the company says.

Tianhe-1A was designed by the National University of Defense Technology (NUDT) in China. The system is housed at National Supercomputer Center in Tianjin and is already fully operational. It will be operated as an open access system to use for large scale scientific computations ...

read more here ...

http://blogs.nature.com/news/thegreatbeyond/2010/10/china_will_claim_supercomputin.html
http://blogs.nature.com/news/thegreatbeyond/2010/10/china_will_claim_supercomputin.html

You're right. This number does have an unusually high failure rate of about 10% on Windows but not on Linux, and it's not the usual memory error. These larger numbers are pushing the boundaries of the siever. The good news is that the failures mostly happen after a few minutes of runtime so not much time is lost and that I have an explicit list of test cases to use in debugging. The bad news is that my plate is full right now, so it will be a few weeks before I can get to it. Thanks for letting me know!

P = NP ? is it bad news for the power of computing ?

Has the biggest question in computer science been solved? On 6 August 2010, Vinay Deolalikar, a mathematician at Hewlett-Packard Labs in Palo Alto, California, sent out draft copies of a paper titled simply "P = NP".

This terse assertion could have profound implications for the ability of computers to solve many kinds of problem. It also answers one of the Clay Mathematics Institute's seven Millennium Prize, so if it turns out to be correct Deolalikar will have earned himself a prize of $1 million.

The P versus NP question concerns the speed at which a computer can accomplish a task such as factorising a number. Some tasks can be completed reasonably quickly ¨C in technical terms, the running time is proportional to a polynomial function of the input size ¨C and these tasks are in class P.

read more ...

New Scientist Physics & Math

Ah yes, that P!=NP "proof" was debunked back in August. :) There are a bunch of posts on a blog by Richard Lipton about it, along with hundreds of comments discussing the issues with the paper: http://rjlipton.wordpress.com/2010/08/15/the-p%e2%89%a0np-proof-is-one-week-old/