Salt Lake City is abuzz with news that China's NUDT team has once again snared the LINPACK benchmark crown at a student cluster-building competition. The team's record-breaking score of 3.014 TFLOPS topped all other competitors and marked the first time a student cluster team has broken through the 3 TFLOPS barrier. This is the …
Moores law says you get a doubling every 18 months
So 5 years of progress should make the systems slightly over 10 times faster, so they are behind schedule on that (about 4GFLOPS looked to be the target). Since 2009 they should be 4 times faster, so about 2.8GFLOPS, so slightly ahead of schedule there.
Re: Moores law says you get a doubling every 18 months
Nope, Moores (original) "law" was the number of components (or tansistors) doubling every year (and not computing power), when he realised his prediction was wrong (10 years later), he changed it do doubling every two years, which of course (because of hindsight) was moore accurate (pun intended), it's now drifting again and approaching every three years.
Moore is an accidental hero, never really that accurate, but held up by many as a visionary, cant really be blamed for that, not his fault, to answer El Reg's question, yes it fits in with the law, if you want it to.
If you want to look to a real visionary, Alan Turing, he made proper verifiable predictions regarding memory capacity, and he was pretty accurate, he predicted computers would have 128Mb memory by the year 2000, which, given the prediction was made 50 years before, was amazingly accurate.
The question is slightly disingenuous because of the power limit. If that has been constant at 26 amps throughout the history then you need significant changes in architecture for boosts. Might be better to plot FLOPs/Watt over time for a better comparison. If the cap has been in place all the time and the improvements have been gained by the move to GPUs then Intel should start to worry.