Interesting Speculation on the Evolution of Processors

[CTho9305]

Originally posted by: Foxery

Originally posted by: pm

The other approach that's not too improbable based on where we are currently is connecting everything with optical interconnects. Companies have demo'd silicon LED's, laser diodes and receivers.

This. Quantum computing is a long ways off, but I believe the next revolution in CPUs is most likely optical signals.

This thread left out one of the major roadblocks in ramping up processor speed: heat. Modern CPUs can't be pushed beyond 4GHz because of a combination of electrical leakage (tunneling) and the inability to keep the chip cold enough. Light has neither of these problems.

See my earlier post. The paper I referenced is this one.

AnandTech recently grabbed an unlocked Intel processor and graphed both the power draw and heat output over several overclocked frequencies. (Great article) Page 2 shows how both curve upwards quickly after 3GHz, and depicts why getting current 45nm processors past 4GHz is nearly impossible without exotic cooling solutions. Skip to page 11 for some eye-openning graphs as well.

I don't think that's a valid analysis. It's sort of like taking a Civic engine and running it in F1 conditions, then saying going 200mph in a car is impossible. Engineers working on chips make many trade-offs that may be the "wrong" choice if you try to operate the chip beyond the intended conditions. Each chip has a power-performance curve and there's an optimal operating range. Heck, take a Silverthorne and try to run it at 3GHz - the power would be through the roof. Or, take a Core 2 Quad and underclock it to run in a cell phone.

 

[imported_wired247]

Even though the process is "45nm" that is object-to-object resolution. There are actually much smaller features on board, namely the gate oxide. The gate oxide was already getting WAY too thin with the old silicon dioxide process, (on the order of 1-15nm, I don't remember) that was in the 90nm and 65nm days. when the gate oxide is too thin, it gives you too much leakage current which gives you increased power consumption that does you no good.

In order to make the leap to 45nm, intel had to introduce a better insulator for the gate oxide, namely hafnium oxide. this is not as easy as it sounds. it changes the entire production process. since hafnium oxide insulates against electron tunneling better than silicon dioxide, they can get it a little bit thinner. But you will doubtfully see the process technology get down to 1nm. the gate oxide would be sub-atomic, which is probably not possible

Another innovation was to use a metal gate instead of the traditional polysilicon gate. I forget exactly what this does, but it results in lower voltages and less power consumption and higher allowable frequencies, IIRC.


eventually, the only way to go is base-3 architecture which is the basis for quantum computing. Now this kind of tech is easily 50 years away, so don't get your hopes up. but it goes something like this...

traditional: 0, 1 (binary)
0=no
1=yes

quantum computing possible states: 0, 1, 2 (base 3)
0=no
1=yes
2=maybe

this would change the entire architecture of the entire computer. It would be like starting the 1960's all over again, except things would be much smaller and faster.

 

[SsupernovaE]

It will be called chromoTronics. Mark my words.

 

[VirtualLarry]

I think that CPUs will evolve to become like modern GPUs - they will have arrays of processing units (functional execution units), and a master thread controller that allows massive amounts of threads to utilize those processing units at the same time. Aka. Massive Multithreading. Think HyperThreading on crack, with a much larger pool of execution resources.