Intel admit Moore's law...

[ALXAndy]

http://www.bit-tech.net/news/hardware/2013/11/22/intel-moores-law/1

It's getting really hard....

 

[stuie]

Was always going to happen theres only so small you can go with current technology, silicon and copper interconnects. The industry has know about these problems for years unfortunately nothing has come along as yet to fix it whats the expected limit 5nm or something and we are already looming on 14 if something big doesnt come along in the next 5 years cpu advancements will come from increasing the size of the blocks but that will start putting the price up

 

[dmferrett]

Only so far on Silicon, but the future is a carbon nano processor. Research is accelerating and looks to replace Si in the next 10 years.

 

[Judgeneo]

Its ok, graphene isn't too far away.


and the UK is leading the way!

http://invezz.com/news/equities/697...-price-more-than-doubles-two-days-after-float

 

[almighty15]

Point is that if all the available software actually took adavantage of all the cores available then we wouldn't really need continuous advancment in terms of CPU's.

 

[dmferrett]

Point is that if all the available software actually took adavantage of all the cores available then we wouldn't really need continuous advancment in terms of CPU's.

Software pushes the necessity for faster and faster hardware. If we want our games and apps to increase in complexity, resolution and detail then the cpu's,gpu'a and ram need to pack more processing power. Clock speeds burn energy whereas more cpu cores use more Silicon.

So unless the process node is smaller the core count probably wont increase at the same price point because cpu silicon is very expensive.

Graphene looks like its the future...

Will we ever see an energy efficient 5Ghz CPU?
Is multicore the future above all else?

 

[ALXAndy]

Point is that if all the available software actually took adavantage of all the cores available then we wouldn't really need continuous advancment in terms of CPU's.

The only reason for Haswell was laptops and tablets. It focused mainly on a better IGPU. There was no reason at all to switch from Ivy to Haswell but people did it any way.

I don't know if they thought that Intel had designed Haswell for them or for what reason you would change but hey, it worked. I would bet the numbers were much smaller than those doing the pointless upgrade from Sandy to Ivy though as Haswell came with problems (heat mainly).

If Intel really wanted to it could push the software developers to work harder at using more cores. But Intel don't want that as their mainstream CPU is quad core. Why spend money pushing people to use eight cores when you only offer four plus HT as your highest spec desktop CPU?

 

[Gazb]

The only reason for Haswell was laptops and tablets. It focused mainly on a better IGPU. There was no reason at all to switch from Ivy to Haswell but people did it any way.

I don't know if they thought that Intel had designed Haswell for them or for what reason you would change but hey, it worked. I would bet the numbers were much smaller than those doing the pointless upgrade from Sandy to Ivy though as Haswell came with problems (heat mainly).

If Intel really wanted to it could push the software developers to work harder at using more cores. But Intel don't want that as their mainstream CPU is quad core. Why spend money pushing people to use eight cores when you only offer four plus HT as your highest spec desktop CPU?

Ivybridge-E is Intel's highest spec Desktop CPU, it's just they class it as enthusiast rather than mainstream. Then they price it out of reach for most people to consider.
Plus you also have server CPU's to consider so Intel does have incentive to push for software to use more than 4 cores/8 threads.
It is going to be interesting to see if the Haswell refresh does improve the IPC or if it will be mainly fixing any issues they'd had with Haswell originally and making it more power efficient.