Devils Canyon at 4.5GHz on passive cooling, how far can DC go?

If they didn't change the TIM then temps wouldn't be lower.
The type of TIM is an upgrade version known as next-generation polymer or NGPTIM.
 
If they didn't change the TIM then temps wouldn't be lower.
The type of TIM is an upgrade version known as next-generation polymer or NGPTIM.

I don't think you saw my earlier comment but the IHS is thinner so it is better for heat transfer.
 
I don't think you saw my earlier comment but the IHS is thinner so it is better for heat transfer.

I saw it, but the TIM is still a upgraded version of what they were using on the previous chips. The thinner IHS helps but the upgraded TIM is also a main contributor.
 
I saw it, but the TIM is still a upgraded version of what they were using on the previous chips. The thinner IHS helps but the upgraded TIM is also a main contributor.

Well I'm only going off of what the worlds current top overclocker is saying so if he's wrong I guess your right as you must have it right infront of you ;)
 
Well I'm only going off of what the worlds current top overclocker is saying so if he's wrong I guess your right as you must have it right infront of you ;)

Well unless he has a scientific lab infront him and has tested the TIM to see it's chemical structure to verify the that it is infact the same TIM, then i'm going to believe that it is indeed NGPTIM.

If he wants to go ahead and remove the TIM from a 4790K and put the TIM from a 4770K on there, then test temps to back up his claims then fair enough.

Until then i'll go off common sense and believe that the lower temps are a result of both the upgraded TIM and a thinner IHS. :)
 
Last edited:
Either way I don't really see the point in this release, Would of been better to just put those resources towards the next REAL performance jump.
 
Another disappointing review, This time from Guru3D -

I think your maximum OC target will be roughly 4600 ~ 4700 MHz if you search 100% stability. Our 4800 MHz tweak was stable enough, but not stable enough for hours of 100% CPU load. And at that stage we already needed 1.4 Volts. I am still running some overclock tests and right now I am at 1.45 Volts / 4800 MHz. This for now is 100% stable however temps are growing to roughly 85 Degrees C.
Review -

http://www.guru3d.com/articles_pages/core_i7_4790k_processor_review,1.html
 
Lol, that "new polymer TIM" thing performs equal if not worse than a normal ( not liquid metal ) thermal paste after delidding...

They should have used their already existing fluxless soldering machines instead of wasting money on this polymer IMHO
 
lol, try putting 1.4v through a 4770k and not having it melt. The 4790k has 1.4v going through it and temps are still ok.
 
lol, try putting 1.4v through a 4770k and not having it melt. The 4790k has 1.4v going through it and temps are still ok.


A conventional thermal paste on a delidded 4770K can lower overall load temperatures by ~10 degrees and with CLU/CLP you can hit 20 degrees less peak temp.
Tomshw test with both cpus @ 4.2 GHz and 1.275V shows that there's just 6C of deltaT between 4770k and 4790k load temps

Moreover I'd never advise someone to keep a 22nm cpu at 1.4V for a daily OC.
 
A conventional thermal paste on a delidded 4770K can lower overall load temperatures by ~10 degrees and with CLU/CLP you can hit 20 degrees less peak temp.

This is with a delidded CPU though, i'm not saying the thermal paste Intel use is better than other thermal paste you can buy. Just that the stuff they are using on the 4790k is better than what they were using on the 4770k. :)

Toms test with both cpus @ 4.2 GHz and 1.275V shows that there's just 6C of deltaT between 4770k and 4790k load temps

Moreover I'd never advise someone to keep a 22nm cpu at 1.4V for a daily OC.

This is at a relatively average voltage though, as you increase the voltage the difference in delta would be larger. Although 1.4v isn't a safe 24/7 voltage it shows how much cooler the 4790k runs at higher voltages compared to the 4770k. If you put 1.4v through a 4770k it wouldn't stay under 85*c and would go into the 90's.
 
Last edited:
Either way I don't really see the point in this release, Would of been better to just put those resources towards the next REAL performance jump.

Im now of the belief that this release was done just to sale some backstock 4770K's. They're the exact same die and a guy on HardOCP identified the date code on the ES as being built in August of last year! I think they slapped some extra caps on the back and maybe tinkered with the TIM a little and called it "next generation", built in a lot of hype and hoped they'd clear out some slow moving inventory. Makes me think they don't pay attention to this community at all and forgot about all these review sites and how the truth would get out about these things before they were released.

Still, as a refresh, and that's all DC is, it's still a great chip. So it won't hit 5.0, that does suck but a 4790K running at a stable 4.7 GHz is still a whole lot of CPU horsepower and still worth buying if you're in the market for a Haswell level upgrade. And considering it's not supposed to cost any more than a few dollars more than a 4770K, I see no real harm done other than to Intel's marketing department.
 
Last edited:
Im now of the belief that this release was done just to sale some backstock 4770K's. They're the exact same die and a guy on HardOCP identified the date code on the ES as being built in August of last year!
I'd be surprized if they weren't. :p

In the industry you don't tape-out new die if you don't have to. New mask-sets aren't cheap. Intel's in a much better position since they don't have to pay TSMC's margins but it still is a high cost for them. It's the reason you've the same Haswell die in everything to low end entry level laptop right up to the flagship desktop 4770k. AMD is the same too.
 
Back
Top