This article explains why Moore's law will stop working around 2018, at the current rate of progress. Intel researchers estimate that by then logic gates will be so small that electrons will be able to tunnel their way through them 50% of the time, making them effectively useless.

http://www.msnbc.com/news/999894.asp?cp1=1

Nasor,

Do you know anything about quantum computers?or anything about qubit concept?well moore's law is here to say more than the time you mention.


bye!

Originally posted by zion
Nasor,

Do you know anything about quantum computers?or anything about qubit concept?well moore's law is here to say more than the time you mention.


bye!

Quantum computing will be a large leap forwards and won't abide by Moore's law. Silicon based chips will become impossible to advance further in the near future.

I think the Intel researchers make a pretty good estimate for known and grown CMOS based technology.

If a radical different concept like quantum computing can be extended, than in fact, Moore's Law wil also end, because it would be much too conservative :p

The current situation is that we see that chipmanufacturers want to postpone the size limits of CMOS by making multicore processors and large L2 / L3 caches and improving PCI / Hypertransport busspeeds instead. But at some point they would reach the limit of current technology and must make a risky investment in new technology.

Postponing the end of CMOS gives intel and AMD the time to let concepts like quantum computing mature a little before they are forced to use it....

Even if Moores Law would end in 2018 wouldn't find that very alarming as we would have tremendous computing power by than...

Moore's law has to do with the number of transistors on a chip. Since quantum computers won't use transistors, they aren't really relevant to the discussion.

Also, keep in mind that quantum computers won't necessarily be faster than conventional computers for all applications. The major benefit of quantum computers is that they can perform an operation on a huge block of numbers simultaneously. An 8-bit quantum computer, for example, will be able to search all possible 2^8 bit sequences for the solution to whatever problem you are solving. This makes them great for things like factoring numbers, but they will probably suck at things like modeling the behavior of systems, where you have to perform a long series of calculations, each of which is based on the result of previous calculations. Since most supercomputers nowadays are used for that sort of linear calculation series, quantum computers won't be replacing ICs any time soon.