Posted by Roland Piquepaille on February 12, 2003 at 07:07:28:
Good morning,
The following is the introduction to a column I wrote today for my weblog ((http://radio.weblogs.com/0105910/).
Best regards.
Roland.
For sure, commodity microprocessors are cheap, but certain tasks still need an extraordinary amount of computing power.
This is what Claire Tristram reminds us with this extremely interesting Technology Review article (http://www.newsfactor.com/perl/story/20730.html) about the Earth Simulator built by NEC for a cost of at least $350 million. Please read it carefully because it's full of good information.
Check this column(http://radio.weblogs.com/0105910/2003/02/12.html) for a summary. In it, I focused on two points: the reason why we still need such supercomputers; and the differences in programming for supercomputers and superclusters. Not only programming for clusters is difficult thus expensive, it is not very efficient.
"The Earth Simulator can crunch computations at up to 67 percent of its peak capacity over a sustained period. In comparison, the massively parallel approach -- well, it doesn't compare."
"What is rarely reported publicly about these systems is that their sustained performance is frequently below five percent of peak, or even one percent of peak," says Thomas Sterling, faculty associate at the Center for Advanced Computing Research at Caltech.