i7 d0 @ 4.0 / 4.2 - temps on a D14?

Ste

Ste

Ste

Soldato
Joined
18 Oct 2002
Posts
2,814
Hi,

Just wondering what peoples opinions on this are. I have an i7 d0 on a udr3. Cooling with a Noctua D14 with the ULNAs (resistors - low noise mode) and a Raven RV02 which has good airflow, fans all on low. I used MX3 paste.


Temps are as follows after 30 minutes small FFTs:

4.0 (191x21) @ 1.248v -> 75/73/72/68

4.2 (200x21) @ 1.31v -> 85/84/80/78 (draws an extra 50w out of the wall, too!)

Do these temps seem reasonable given the set up I have or is it likely I could squeeze out a bit more performance with a reseat/reapplication of thermal paste? What do people think of MX3 vs AS5 as I have both lying around!


Couple of other things whilst I'm here:

- Does QPI/Vtt have much effect on temps? Testing 4.2 was quick and dirty and I just put QPI straight up to 1.335 - is it worth trying to get it lower, will that help temps?
- I am thinking that 4.2 clearly isn't viable at the moment as I guess IBT et al would show 90'c plus - no thanks. That said, what do people think of running 85'c on P95? Too hot or okay?

My idle temps are about 41/37/40/35. Given that the summer is coming I think that I need to shave at least 5'c load off at 4.2 for it to be a sustainable overclock, do you think that is possible or am I just asking too much on air?

Many thanks for any pointers :)
 
Hi your CPU - Vcore in bios is about right, mine are that at 4.0GHz / 4.2GHz, slowly lower your QPIv till it affect's the OC go back up one and test(hows your case air flow?).

If l remember right you can set the Noctua fan's to > low / high RPM, if it was me l would set the fan's to high RPM as the noise level:13.2dB(A) should be silent, which will bring your temp's down.

I use MX3[no curing] but AS5 is still very good, you want to run P95 for at least 6 to 8HRs, some will say 12HRs, to test for stability.

Using a H50 x1 Akasa Viper x1 Apache push pull config, P95@small/fft's cores: 72'c:74'c:74'c:76'c after 2hrs. I had already tested for stability at 4.2 using to Apache fan's[see:sig]
 
I agree with Oldphart, and if it is to noisey for u then just use speedfan to automate the speed according to temps.

Trying to get the system stable with the lowest possible Qpi/vtt is key, as that will cause a lot of heat, if to high. I have both my vcore and qpi at 1.275 in bios for my clock in my sig, with load line calibration on that eqautes to 1.26v in windows

About the thermal pastes both are good, personally id go for Mx3 as no curing time needed, so u know the performance is there from the get go.

Not bad temps at all to say u had the fans on low, think when i have my vipers running low, my highest core temps are high 60's, but then again i am on custom water lol.
 
I'd do the same, get your fans RPM up then re-test, Noctuas aren't very noisy apparently, never used one myself. But as your OC'ing 4Ghz+ low RPM isn't the best idea. Not unless your water cooling then low rpm is ok :)
 
Cheers guys. Might give that a go, though the reason I went for low RPM was that there was a 1-2'c difference max at 4ghz. Perhaps it's a different story as the temperatures go up?

I'll give it a go and report back - nothing with those temps that makes anyone think I should be reseating the heatsink etc. though?

Cheers
 
There is a clear audible difference with Noctua fans pushing air through a Noctua heatsink when the fans are slowed down to when they go up to full speed.

I have further reduced the speed of my fans by using Speedfan to slow them down to near a inaudible 600rpm, then speeding up when needed to 1000rpm, then there is a clear difference in noise coming from my case. I run a slightly higher idle temperature as a result of course but that's better than the noise.

I suppose it all depends on the rest of your system and what level of noise it produces. If you have a silent system then you will hear it, if not then blast away.
 
Yep there is, even with the noctua resistors in it's the loudest thing in my case - people have different definitions of silent!

With the fans cranked and a couple of notches down on the qpi volts I'm at 80 hottest core in prime. Getting there, any other ideas? The heatsink fins are warm to the touch so the heatpipes are doing their job. Thanks
 
those temps seem a bit high unless the D14 loses to the venomous X with a single fan and some old silver TIM by a good 5C on the first 3 cores which I doubt
 
HW91 said:
those temps seem a bit high unless the D14 loses to the venomous X with a single fan and some old silver TIM by a good 5C on the first 3 cores which I doubt
Click to expand...

They are very close tho, i think in helios's review the noctua is ahead by only 0.5oC , the VX are very good considering the size compared to the noctua..
 
I hear that i7 920s differ somewhat but that's a very high clock and voltage for such a low temp. Mine goes to the high 80s at 1.25 but I'm not too bothered. The highest I've ever seen it go under gaming is mid 60s and mid 70s in 3dsmax rendering.

I'm only using the 120mm Noctua variant though.
 
you need to be expecting around 80'c at 4.2GHz

i have been getting around 84'c in our technical department when we tested the cooler

unfortunatly due to its size and weight it means we can use it for system integration as it poses a courier damage risk.

i have had a lot of people customise bundles to include one though :)
 
Thanks - what voltage is that at? I managed to get linx under 85c but it isn't stable after about half hour :( I guess I'm fairly close temp wise as I'm in a sotuh facing room and it gets hot in there!
 
You must log in or register to reply here.
Back
Top Bottom