Cpu physics, speed, posabilities

[Shadow_Storm56]

How fast is too fast? is  8ghz, 15ghz, 35ghz even possible ever? or would this not be possible unless cpus were made completely different? or is it more likley for us to see 10, 20 , 100, 1000 cores and advanced multi threading at more current speeds?

[iRileyx]

There'll never be a processor that's too fast. as long as all the other components can keep up with it.

[Bajantechnician]

3 minutes ago, Shadow_Storm56 said:

How fast is too fast? is  8ghz, 15ghz, 35ghz even possible ever? or would this not be possible unless cpus were made completely different? or is it more likley for us to see 10, 20 , 100, 1000 cores and advanced multi threading at more current speeds?

 

Just now, RKRiley said:

There'll never be a processor that's too fast. as long as all the other components can keep up with it.

whong. too fast is when there is nothing in the world at all can active cool it so it doesnt blow up  

[Shadow_Storm56]

I mean in the realm of physics is there a point whare there would be too much  heat generated to have a certain clock speed

[Shadow_Storm56]

I just wonder what more likley in 20 years for example.... 20ghz or 30 cores....

[iRileyx]

You should have said that in your first post because now youve mentioned heat it completely changes the question lol.

[iRileyx]

1 minute ago, Shadow_Storm56 said:

I just wonder what more likley in 20 years for example.... 20ghz or 30 cores....

Theres a 22 core Xeon that already exists, but 20ghz seems highly unlikely unless someone makes a breakthrough discovery in cooling.

[SCHISCHKA]

AMD FX has the record at 8GHz. AMD achieved this by creating a CPU that sucks at games

https://en.wikipedia.org/wiki/Clock_rate#Historical_milestones_and_current_records

Quote

As of 2011, the Guinness World Record for the highest CPU clock rate is an overclocked, 8.805 GHz AMD Bulldozer-based FX chip. It surpassed the previous record, a 8.670 GHz AMD FX "Piledriver" chip.^[3]

 

[Justarandomperson]

i believe a 16 bit 1000 core processor already exists(made by stundents and because of the architecture not useable fore a desktop)

something different:

transistors can be very small but at a point they are so small that quantum physics set in and electrons can just "jump" so that would be a limit of making a transistor smaller then cpus just had to grow(just a random tought)

 

[Guest]

Just now, RKRiley said:

Theres a 22 core Xeon that already exists, but 20ghz seems highly unlikely unless someone makes a breakthrough discovery in cooling.

Canada

[Shadow_Storm56]

This is my point. whare do we go and how will it go

[pwn_intended]

Core count is limited only by die size and size of the actual core themselves. I remember reading somewhere that the limitation for clock speed is the speed of light (more accurately the speed of electrons through the chip itself) and that the upper limit is somewhere in the 5-6ghz range. I could be totally off on that one though.

[Shadow_Storm56]

Hmmmmmmm

[iRileyx]

I doubt anyone here will be able to tell your nor predict what cpu's will be like in 20 years, and thats the simplest answer you'll get lol.

[Omon_Ra]

5 minutes ago, Shadow_Storm56 said:

I just wonder what more likley in 20 years for example.... 20ghz or 30 cores....

Quantum computers. We're already creeping towards the end of what is achievable with silicon.

[vorticalbox]

4 minutes ago, Shadow_Storm56 said:

I mean in the realm of physics is there a point whare there would be too much  heat generated to have a certain clock speed

No if we used light and seeing as with can cool materials down to almost absolute zero I think we can handle the heat of a chip.

 

The next step will be moving away from silicon.

[Shadow_Storm56]

It's just a curiosity of what's more likely. 

[Mira Yurizaki]

The problem with processor speed is the power dissipated. So I'm going to pull out this formula: P = CV^2f. This is the amount of power dissipated per transistor. So the entire package can be though of as P = nCV^2f where n is the number of transistors.

 

n and f are the billions. Obviously having 10^18 watts of dissipation is bad news bears. So you have to do your damnedest to get CV^2 as tiny as possible.

 

Good luck with that :3

[EunSoo]

7 minutes ago, Shadow_Storm56 said:

I mean in the realm of physics is there a point whare there would be too much  heat generated to have a certain clock speed

We will probably have created a completely different architecture that will allow us to run at that speeds

[Shadow_Storm56]

Just now, vorticalbox said:

No if we used light and seeing as with can cool materials down to almost absolute zero I think we can handle the heat of a chip.

 

The next step will be moving away from silicon.

Thus my point so the prediction is silicon will be retired and new ways will come in..? prolly

[Shadow_Storm56]

Or maybe we will go for more cores instead of more speed?

[Sakkura]

9 minutes ago, Shadow_Storm56 said:

I just wonder what more likley in 20 years for example.... 20ghz or 30 cores....

Unknown, since we will have to move on from silicon before then. But off the shelf silicon CPUs will likely not get much past 5 GHz (with normal cooling, of course LN2 overclocking can do crazy clocks).

4 minutes ago, Omon_Ra said:

Quantum computers. We're already creeping towards the end of what is achievable with silicon.

Not necessarily, as quantum computers aren't faster at everything.

[Shadow_Storm56]

My prediction is that we will have a completely differvent way of computing, somthing we can't even imagine. 

[pwn_intended]

4 minutes ago, Shadow_Storm56 said:

Or maybe we will go for more cores instead of more speed?

Well that is exactly what has been going on for the last decade.

[Shadow_Storm56]

Maybe instead of quad core ull have 4 cpus running at 6 or 7ghz but with only 1 core sof you can dissipate heat