And still they squeeze more and more processing power out of silicon!

[Neil]

http://www.bbc.co.uk/news/technology-13283882

[brer]

It's a stopgap IMHO. Moores law if it has not broken down yet, will do so in the near future.

What they are attempting to do is layer the circuitry and move on to the next smaller size. Quantum effects will end up dictating how small a transistor can be made. Layering the circuitry gives them more raw power in a smaller space, but will eventually cause problems with heat.

One of the big rules about computers from the bad old days of relays and tubes was that the maximum size of a computer was determined by how much heat it could dump.

When we made the shift to intergrated circuits, we sidestepped the rule but did not avoid it. Linking computers together also gives hope for a few more years of increasing power, but that will eventually reach a maximum size based on cost.

I think a lot of the optimizations in the future will be based on increasing the efficiency of the base code that the operating system runs on.

I am not a hacker, but even I can write a good routine in assembly that will run far faster than one written in C or C++. Writing the base code more efficiently will likely give us at least another decade of increasingly powerful computers while retaining small sizes.

Yup, I'm a happy camper. I just finished writing the assembly code to drive a small cpu to control a well pump from a problem well. All I have to do is proto the control board, burn the chip, and run a little conduit.

Me, I actually like the smaller processors a lot more.

[Neil]

You'd think at some point in the next ten years a radical change will come, moving us away from traditional silicon processors... And that no doubt will move us into processing speeds far greater than today... Possibly even enough to run Doom4!

[BillyRay]

http://www.bbc.co.uk/news/technology-13283882

Why am I in the mood for waffles, now?

[brer]

I do not think we will move from silicon anytime soon. It has too much going for it from a material standpoint.

It can handle fairly high temperatures for one. It it easily purified into a decent semiconductor material for another. It is plentiful for another.

Photo lithography was turned into a high science in its own right because of silicon.

I am not saying that we are stuck with silicon, just that all the alternatives are decades away and are all more expensive.

What is kicking us in the head is not the material, but basic quantum physics.

Materials at a nano level do not respond the same way as they do at a macro level. This concept becomes more and more important as we keep trying to make the circuitry smaller.

At a macro level, entropy exists. Heat goes from a warm object to a cold object. An object will roll down hill. Electricity will only flow in one direction based on polarity and potential.

At the micro level, these rules break down as if they are based on statistics instead of a hard rule.

Even changing the materials we use as semiconductors will not buy us much more time in this regard.

Architecture and operating systems will, in my opinion, be the next things that are optimized in order to provide succeeding generations of devices. Devices will include more processors designed around parallel computing as will the operating system.

But on the other hand, just wait until we start getting the first generations of computers designed around using some of the more esoteric effects of quantum theory. Memory buses that communicate via particle entanglement, quantum processors based on electrons using undefined states or a few of the other tricks that are being developed.

They will be cool as heck when they finally arrive, but this may be a few decades away.