• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Intel working hard on 22nm CPUs

The future does look bright.

Definitely waiting for 32nm before I switch over to i7 if its coming out in 2009. Hopefully everything should be cheaper (mobo/memory) when 32nm is launched.
 
There seems to be this barrier in clock speeds because of heat, which has facilitated a rise in multi-core CPUs. I really want Intel to be trying to achieve higher clocked processors, because softwareis probably around 5+ years behind the hardware trends.

And I though this new non-silicon material would allow for higher clock speeds, but it seems not.
 
A smaller manufacturing process generally means less power consumption and less excess heat on the same architecture (for eg the move from 65nm to 45nm) on which higher clocks are attainable due to this fact, if Intel decide to use up that extra allowance for something more beneficial then I'm all for it, which initially they wont as the 32nm variant of the 45nm i7 will just be the same architecture on a smaller process, possibly with a few fixes or extra features but using the same basic architecture (tick tock)

http://www.intel.com/technology/tick-tock/index.htm
 
it's not as clear cut as that. it was never the heat the cpu's as a package produced that was the problem per se from my understanding, but rather a problem of switching those transistors on and off reliably at a rate of 5 billion times a second (5ghz). they have to deal with current leakage which is a result of power input, process size and and switching frequency. The faster they switch the more power they require to do so and the more power they leak. There's only so much they can do to combat that and making the process smaller only makes this problem worse because the current leakage vs power input is increased. Intel didnt know this, untill they tried to scale the pentium 4 netburst architecture. They famously said they expected netburst to scale to 10ghz when it topped out at 4 because of inefficiancies of the design, both in speed and power usage. Thats why they abandoned netburst and went back to developing the core technology.
 
Last edited:
Duff-man said it, before you hit the limit being the size of the atoms you hit quantum effects. They will be able to get a bit smaller with gallium and indium, as they're larger so tunnelling is reduced. However these will be much more expensive to produce, as they aren't as abundant, even though the technology to produce them is basically the same.
 
it's not as clear cut as that. it was never the heat the cpu's as a package produced that was the problem per se from my understanding, but rather a problem of switching those transistors on and off reliably at a rate of 5 billion times a second (5ghz). they have to deal with current leakage which is a result of power input, process size and and switching frequency. The faster they switch the more power they require to do so and the more power they leak. There's only so much they can do to combat that and making the process smaller only makes this problem worse because the current leakage vs power input is increased. Intel didnt know this, untill they tried to scale the pentium 4 netburst architecture. They famously said they expected netburst to scale to 10ghz when it topped out at 4 because of inefficiancies of the design, both in speed and power usage. Thats why they abandoned netburst and went back to developing the core technology.


Netburst would have burst into flames at 10GHz!


22nm and even greener computing as temps will be cooler
 
Some parts of the faster P4s were running at upto 7.6GHz at stock speeds and well over 10GHz when overclocked.

We're thinking of the ALUs which were clocked twice the speed of the rest of the core right? I suppose 16GHz for the ALUs on a Cedar Mill under liquid nitrogen isn't so bad.
 
Aye, twice the speed and as you said the record for P4s is just over 8GHz, so as you say 16GHz in the ALU. Wouldn't have been stable though :p
 
I doubt anyone would complain if their Quad core was running at 8gHz :p

22nm is just the future, it's not going to be any more revolutionary than the previous changes in process.
 
hopefully by the time these 22nm processors come out then SSD hdds should be cheaper larger and faster. If thats the case then teh performance difference for a computer running a 22nm processor - (I asume) DDR 4 ram the latest SSD and the latest graphics card should be one of the biggest we have ever seen.
The performance diffrence should be a lot bigger than going from the P4 netburst to the Core2

That is all IMHO of course
 
always thinking ahead

I saw a documentary a while back saying that they may move to an organic based architecture.

So I started stockpiling wood in my garage which I plan on selling to intel for the right price. :D
 
why buy wood? all intel would need is provide free pr0n for their male engineers/staff in general and there's plenty of organic material just lying around for the taking :eek:
 
This is very interesting and hopefully Intel will be able to go there, it seems that the graphics card industry is going nowhere because of power leakage and they are having some serious problems on their hands, if Intel can produce Nehalem & SSDs using a working 22nm process that would be real progress.
 
Back
Top Bottom