Intel Devil Canyon officially unveiled

[royfrosty]

I wonder what voltage and cooling they are using......

According to their cheque price it states air/water

 

[allhomme]

Judging from how much Haswell fluctuated in clocking for the peasants chips, I think we should adopt a wait and see approach to Devils Canyon.

 

[dawnknight]

Huh......... SB/IB/Haswell all can run 4GHz at stock voltage using stock heatsink.......
No hassle needed even..... Just set multiplier to 40x in BIOS..

oh...what u described here, technically, is what is now predominantly being described as "overclocking" lol so your reference frame point of view is somewhat different

Lol..... I guess we are hitting limits of current silicon technology. Despite small process, it does not aid overclocking at all....I was so disappointed with the Tri-gate and 22nm process when IB came out. It does not improve oc but actually degrades it....lol...

...
Most if not all "overclockers" now are just targeting the 5GHz mark to boost their e-peen. They wont even know if there will be a noticeable performance improvement between 4.2/4.4GHz and 5GHz, a paltry 0.6GHz.

from core2quad to sandybridge (sb), the cpu speed increase @ load is about 2 to 3 times. this observation, i have verified it via experiments on video encoding, and mathematical/scientific computation-intensive programs.

on sb, overclocking from stock 3.4ghz to 4ghz, the speed increase @ load is about 15%. (verified via encoding and computation programs)

then from sb to devils canyon, according to one of the released images comparing 2700k to devil's canyon, intel claims its about 1/3 increase w.r.t sb. this claim, i have yet to verify lol.

but my sense is similar, i do somewhat agree with the general sense that the supposed regular/annual advancements (intel tick-tock approach) appears to be approaching an asymptote i.e. apparently hitting some kind of limits, silicon tech-wise. apparently the inherent inner-workings of quantum mechanics still rules and imposes some supposedly uncrossable barriers, despite human engineering efforts to continuously maximise the benefits out of it. or maybe we are not that advanced to cross those barriers yet. so given the existing status quo of still employing semicon-based computing tech, pretty much soon the wall will be hit. would be interesting to see how the "hitting the wall" scene will be like in future years lol

 

[Yureak]

from core2quad to sandybridge (sb), the cpu speed increase @ load is about 2 to 3 times. this observation, i have verified it via experiments on video encoding, and mathematical/scientific computation-intensive programs.

on sb, overclocking from stock 3.4ghz to 4ghz, the speed increase @ load is about 15%. (verified via encoding and computation programs)

then from sb to devils canyon, according to one of the released images comparing 2700k to devil's canyon, intel claims its about 1/3 increase w.r.t sb. this claim, i have yet to verify lol.

but my sense is similar, i do somewhat agree with the general sense that the supposed regular/annual advancements (intel tick-tock approach) appears to be approaching an asymptote i.e. apparently hitting some kind of limits, silicon tech-wise. apparently the inherent inner-workings of quantum mechanics still rules and imposes some supposedly uncrossable barriers, despite human engineering efforts to continuously maximise the benefits out of it. or maybe we are not that advanced to cross those barriers yet. so given the existing status quo of still employing semicon-based computing tech, pretty much soon the wall will be hit. would be interesting to see how the "hitting the wall" scene will be like in future years lol

The issue isnt technology... Honestly theres no motivation for intel to keep to moore's law since amd is so shitty

 

[Yureak]

Devil's Canyon still cannot annihilate the prowess and might of Bulldozer at 8 GHz mark...

Except a bulldozer at 8ghz is just a fireplace heater.... Calculations? What calculations

 

[Yureak]

thanks for pointing that out.

that means whatever info that the mass public receives and knows about real silicon tech, has been skewed by a layer of inherently non-technical issue i.e. power struggle/amd vs intel/profits/financial issues stuff etc

however, with or without this additional layer to skew/mask the true nature of advancement/limits of silicon tech, i.e. whether u look at a skewed or non-skewed silicon tech advancement curve, my sense is that it will still reach some asymptote eventually, i.e. the laws of quantum mechanics will eventually break moore's law. and when that day happens, i wonder how the world, in those future years, be it consumer sentiment, intel's response, etc, future advancements, will be like lol

It definitely will. What will happen when node process drops below 10? Im not sure.

Yet what is certain is that Intel vs AMD is playing out like Nvidia vs AMD

Nvidia likewise released a midrange chip as the 680 to charge top end prices since amd didnt have anything that can fight.

Likewise this is how 880 is going to be gimped

 

[manshen]

Except a bulldozer at 8ghz is just a fireplace heater.... Calculations? What calculations

At least it can still run SuperPI

Der8auer Sets New AMD SuperPI 32M Record During Pro OC Cup - CPU Clocked Over 8GHz!

The 6.4GHz devil canyon can't run anything.

 

[Gattberserk]

So 5.5GHz Devil canyon on air/water?


I guess 5 GHz on air is indeed possible. But I am more interested in the voltage it require to stay there permanently.

 

[Yureak]

At least it can still run SuperPI

Der8auer Sets New AMD SuperPI 32M Record During Pro OC Cup - CPU Clocked Over 8GHz!

The 6.4GHz devil canyon can't run anything.

Why cant a devil's canyon run anything?

 

[spyboy48]

Intel® OC Challenge Computex 2014 @ HWBOT

 

[Gattberserk]

I am glad Intel took a little more attention on us rather than just blindly focus on mobile devices.


I need more consumer level review on how well OCing does compared to 4770K. Basically beside better TIM and also more capacitor to deliver smoother power, there is nothing else right? No improvement in performance per clock

 

[devilwahaha]

I am glad Intel took a little more attention on us rather than just blindly focus on mobile devices.


I need more consumer level review on how well OCing does compared to 4770K. Basically beside better TIM and also more capacitor to deliver smoother power, there is nothing else right? No improvement in performance per clock

Its probably just a very small side-step. After all this isnt their money making category.

Its pretty much known to all Intel doesnt really give a **** about "enthusiasts", they are simply at the mercy of Intel and picking up whatever they release.

 

[gaoxing84]

Huh......... SB/IB/Haswell all can run 4GHz at stock voltage using stock heatsink.......

No hassle needed even..... Just set multiplier to 40x in BIOS.

This is not a step forward at all. The only real challenge is hitting 5GHz which is very hard to achieve.

that's true. anything more than 4ghz, usually need to bump up vcore from stock

 

[MacClipper]

I wonder what voltage and cooling they are using......

Do note how some results were obtained eg. so called air or water cooling ...
Overclockers Forums - Post 171
Overclockers Forums - Post 174

The problem is, even the people OCing on H100 and so, are using LN2 before the fans, to draw cool air, and the amateurs were maxing out at 5.1ghz.

All the teams ocing today work for Intel, so i highly doubt they were given the worst in the batches.

... they were using LN2 to pre-chill the intake on the "air" cooler. Or at least Gigabyte was.