CPU how much better can they get?

[kEn_sct]

OK as far as i understand a new cpu is better because it has more transistors.

So they make the transistors smaller and smaller, but how much smaller can they get, before they become 2 small? and when this happens whats going to repace them? is there any compleatly new CPU designs in the works? or are we going to hit a max CPU speed? or are they going to just pack more cores in till it gets ridiculos?

 

[doiks14]

maybe we'll move on to something organic/biological/chemical instead of electrical?

 

[Phatzy]

i heard 11nm is the limit for silicon, they just use other kind of element i bet

 

[kEn_sct]

i heard 11nm is the limit for silicon, they just use other kind of element i bet

well acording to the intenet a silicon atom is 0.234 nm across.. so if thats 11nm ias true that means the transistor would be only 55atoms across!!

 

[Keto]

http://en.wikipedia.org/wiki/Moore's_law
http://en.wikipedia.org/wiki/Moore's_law#Future_trends

 

[Xappy]

On the current tech we can keep adding complexity till we get to the minimum transistor size at the maximum die size for a single clock pulse running through silicon.
Then possibly some way of adding complexity by working a bit in 3 dimensions though there's limits to what can be done.
Then we have to start looking at something more radical.

 

[Ch3m1c4L]

HP or IBM are working on "stacked" cpu's, one chip, but several cpus on top of each other with water running inbetween them to cool them. Theres also quantom processors that have been made now, although i dont understand them and dont think they are useful for our needs.
As stated above, different elements, also different cpu arcitecture etc helps. Chances are well end up with a silly number of cores at some point, which will be fun

 

[madindehead]

The main problem currently, is we can pack millions of transistors onto a chip, but we don't actually know what to do with them! So more and more go into cache instead of being used to process things. I'm sure they will find something. When we hit the size barrier, they will just have to find different CPU architectures. We fill find something. I might even be involved when I finish my electronic engineering degree =D

 

[VortX]

Intel are currently working on cellular CPUs. These are of course in the experimental stage and won't be out for a long, long time. Not to mention the fact they take several years to grow.

On top of that, and will most likely come out first, I remember reading an article on a CPU manufacturer, most probably Intel or IBM, who are working on CPUs that communicate via light, much like fibre optic cables, but in a CPU.

This is because not only can light travel MUCH faster than electrons (It's the fastest thing known in our universe obviously), but it should also generate little to no heat.

Quite a nice little upgrade eh?