bhoolkumehul
Broken In
why intel and AMD stuck in processor speed up to 3.2 to 3.5 GHz? why there is no processor speed more than 4.0 GHz?
mukherjee
Treatment in progress...
@ OP
U see,the proccy die as a whole has set TDP(say 95w).
If there was a single core,it cud use all the TDP,and hence reach higher clock speeds,remember the Pentium 4 EE?
As the no of cores increase,the available TDP is divided among the cores(say quad core,hence TDP per core=95/4=23.75 appx). Hence the speeds they can attain becomes limited(this is the basis of Intel turbo boost and AMD turbocore).
U see,the proccy die as a whole has set TDP(say 95w).
If there was a single core,it cud use all the TDP,and hence reach higher clock speeds,remember the Pentium 4 EE?
As the no of cores increase,the available TDP is divided among the cores(say quad core,hence TDP per core=95/4=23.75 appx). Hence the speeds they can attain becomes limited(this is the basis of Intel turbo boost and AMD turbocore).
coderunknown
Retired Forum Mod
i think its mainly cause of the current node size & market strategy.
with 45nm manufacturing tech, higher clock will increase temperature lot. so will require really good cooling. so as Intel releasing the 2nd gen 32nm batch of Core i processors, the clock is increasing but still the temperature is kept within limit (node size) & we'll see AMD follow Intel in around 6 month.
also Intel showed that a lower clocked Core i beats a higher clocked Phenom II. why? its mainly cause their architecture is more efficient, so no use going for a higher clock & reserving it for future (market strategy). on other hand AMD uses less efficient architecture. so you get more transistor but less performance or almost same performance (at same price).
with 45nm manufacturing tech, higher clock will increase temperature lot. so will require really good cooling. so as Intel releasing the 2nd gen 32nm batch of Core i processors, the clock is increasing but still the temperature is kept within limit (node size) & we'll see AMD follow Intel in around 6 month.
also Intel showed that a lower clocked Core i beats a higher clocked Phenom II. why? its mainly cause their architecture is more efficient, so no use going for a higher clock & reserving it for future (market strategy). on other hand AMD uses less efficient architecture. so you get more transistor but less performance or almost same performance (at same price).
vickybat
I am the night...I am...
I agree with sam.shab's take on this. When a die-shrink occurs, tdp lowers and therefore less heat is generated. So we have higher clockspeeds in the sandybridge cpu's than the bloomfields and lynnfields. And rememeber the stock cooling solution given by these companies are somewhat limited and designed for stock speeds.
So future 22nm or lesses fabricated cores will take the stock speeds beyond 4ghz levels. Architecture design also has a major part to play in this and not alone fabrication methods.
So future 22nm or lesses fabricated cores will take the stock speeds beyond 4ghz levels. Architecture design also has a major part to play in this and not alone fabrication methods.
pulkitpopli2004
Cyborg Agent
^^no its not like that..
designers always try to design a chip that can achieve maximum frequency of operation.. but due to some limitation of CMOS integrated circuits, frequency like 4Ghz in proccy can't be achieved..
n yeah sam.shab is also correct..
designers always try to design a chip that can achieve maximum frequency of operation.. but due to some limitation of CMOS integrated circuits, frequency like 4Ghz in proccy can't be achieved..
n yeah sam.shab is also correct..
^^ but if you can OC some cpus will easily pass 4GHz mark and SB pus can reach 5Ghz - so there's no such limitation.
I think it's a pure market strategy :
manufacturers don't release cpus with maximum clock speed to keep enthusiasts users happy and - it's upon to the users to take their cpus to the maximum clock speed possible and it really feels great if you can get 1-1.5 Ghz more than the stock clock speed and we all know to achieve higher clock speed we should have a decent mobo and enthusiast users don't mind in paying for quality mobos to achieve the maximum possible OC and this keeps the enthusiast users happy and boosts the sales of highend mobos at the same time.
Another thing is to get higher clock speed you need a better cooler as well - so it's even good for 3rd party cpu hsf makers and these points are even true for gpu market.
Also not every cpu and gpu can be oced at the same same level - users have to find the maximum possible stable clock speed - if manufacturers wants to do the job instead they will have to put in more resources and that will increase the cost of everything.
So manufacturers release cpus and gpus with a generic cock speed.
I think it's a pure market strategy :
manufacturers don't release cpus with maximum clock speed to keep enthusiasts users happy and - it's upon to the users to take their cpus to the maximum clock speed possible and it really feels great if you can get 1-1.5 Ghz more than the stock clock speed and we all know to achieve higher clock speed we should have a decent mobo and enthusiast users don't mind in paying for quality mobos to achieve the maximum possible OC and this keeps the enthusiast users happy and boosts the sales of highend mobos at the same time.
Another thing is to get higher clock speed you need a better cooler as well - so it's even good for 3rd party cpu hsf makers and these points are even true for gpu market.
Also not every cpu and gpu can be oced at the same same level - users have to find the maximum possible stable clock speed - if manufacturers wants to do the job instead they will have to put in more resources and that will increase the cost of everything.
So manufacturers release cpus and gpus with a generic cock speed.
..
..
pulkitpopli2004
Cyborg Agent
OCing also has adverse effects on CPU.. sometime can damage that too..
i think thats why stock clock speed of arnd 4Ghz has not been reached yet..
and so dependence is more on manufacturing technology and circuit designing rather than market strategy..
^^ If you are not aiming for some extreme OC to set world record then there's no harm in OCing as long as you use some 3rd party cooler and can keep the temps under acceptable range.
As I said before not every chip is equal and will not OC to the same limit - some will end up at say 3.9 Ghz and others may pass 4 GHz mark - but if manufacturers wants to test out the maximum OC ability of each cpu then it will be a very expensive process for them.
So they just release every cpu in the market with a standard speed.
@ ssb1551 - thanks for your comments.
As I said before not every chip is equal and will not OC to the same limit - some will end up at say 3.9 Ghz and others may pass 4 GHz mark - but if manufacturers wants to test out the maximum OC ability of each cpu then it will be a very expensive process for them.
So they just release every cpu in the market with a standard speed.
@ ssb1551 - thanks for your comments.