What will happen at ultra high GHz frequencies on a CPU

The reason CPUs generally don't run at more than ~5 GHz is a matter of cooling, mostly.
To make the chips run faster, you need to feed them more voltage, which creates more heat. You end up needing crazy cooling setups like refrigeration or LN2 to deal with all that heat being created in such a small footprint.

Engineering improvements and die shrinks have driven most of the increase in clock speed, but after 7nm there is probably only one die shrink left because the transistors are so small that electrons will jump from one to another. We'll probably see another bump in clock rates then, but it won't be huge because of thermal limits.

After the die shrinks, the only real improvements will be in the design to increase instructions per clock. Then, barring the introduction of a new technology (i.e. graphene or optical processors), software will have to actually be optimized! Software devs won't be able to waste tons of processing power on 50 MB "hello world" - tier programs. Parallelism will also become very important, because it probably won't be uncommon to have a computer with 64+ cores.

Wrong board for rp shit. is that way

Kys

It will achieve nuclear fusion, intel is already working on it, they'll market it as self-powering CPUs

Attached: power density.jpg (960x720, 90K)

The picture you posted is wave frequencies, CPUs are not using that, so they are completely unrelated. CPU frequency determines how often impulses of electricity are sent through the circuitry inside CPU, and at frequencies higher than 4-5 GHz the shape of those impulses (they are supposed to be squares) becomes so distorted that shit stops working correctly.

I doubt we'll ever see clocks significantly higher than we got now, e.g. < 5 GHz stock. There perhaps shall be more cores and IPC improvements but all in all the technology is approaching physical limitation, so there either will be something completely new kicking in (which I doubt to see next two decades at least) or we are stuck which is more realistic.
In fact I do not anticipate seeing 5 GHz stock CPU ever.

this.

T. Retard that thinks hertz are exclusive to electromagnetic emissions.
This chart has nothing to do with CPUs, not even remotely.

This thread is one big bait.

>Intel jews reportedly broke the laws of thermodynamics

More false advertising at 10

This guy is right, we need to invent 30 ghz suoper cpus so we can communicate with the spiritrealm whichj also ioperated at 30ghz

We could go higher than that by redesigning transistors, that's what we're doing currently as it is the only way of getting around moore's law

5GHz stock might be possible after the next die shrink and lots of design optimization.
CPU improvements have been pretty slow the past 5 years, and I'd bet the next 10 are going to be even slower.

This is the dumbest shit I've read all week

Make a single core with a long ass pipeline and cramm in enough current to overpower the circuit's resistance to get readable output at the end and we could... Oh wait that's Netburst. Intel was ahead once again

>but power consumption increases according to the square of the frequency
It's square of the voltage. It's linear in frequency.

...

> What will happen when we hit 30GHz on a desktop PC CPU?
It will give you cancer and you will die

Nigger what

>30 Ghz CPU
Is it even practically possible?
Even an atomic clock is 3 times slower (about 9Ghz)

Also, you'd need a tiny CPU. At that frequency, light only travels 1cm (0.393 inch) per clock tick, and the electrons in your system are actually slower than that.