Another i7 Temp Thread? yeaaa..

[Nanometer]

The rig is listed below, and I just remounted my water block to rule out possibly bad IHS contact.

With all the fans on full, My temperature shoots up to 77c when I'm feeding it 1.41 volts in my quest to 4.5 GHz.

On idle I sit at 40c on 23c ambient air(fans on full). I load the CPU on prime and within 2 seconds it shoots up to 72c. I am thinking that my pump might not be strong enough? I was thinking bad IHS, but I'd like it to be able to be something I could fix.. haha

[century child]

The only was your going to get significantly lower temps is by using a water chiller. At the amount of voltage you are pushing to that CPU 77° is fantastic. I think you may be expecting too much.

[Nanometer]

This isn't on air though, I think the temps should be much better then they are.

[SoulsCollective]

With sig-rig, my 920 idles at around mid-50s, and loads in low-70s. Using about 1.35vCore.

Your temps are fine.

[Nanometer]

I see... What is the weak side? heat transfer to the block im guessing?

[gmat]

i7's are producing a LOT of heat. HT is the main offender there, also they have an upgraded microarchitecture with a lot more transistors than previous gen. Your temps and max OC are very good already, far above what most can do... (i reach these temps at 4.3)

[Nanometer]

I have ht off though haha

[Vinas]

Yup, I can tell it's getting warmer outside from all of these temperature threads ;)

It seems like for that speed and vcore that your temperatures are quite good actually. My i7 860 chip can't run past 4.2GHz and at that speed I'm hitting 80's constantly with linpack and cooling in my sig. That said, if you got a newer block you might see a 1-3C improvement. Eddy EK and folks must love this time of year!

[SoulsCollective]

Yup, I can tell it's getting warmer outside from all of these temperature threads ;)...Eddy EK and folks must love this time of year!Where I am, it's heading into winter.

USA =/ everywhere ;)

[Vinas]

Where I am, it's heading into winter.

USA =/ everywhere ;)True that Souls, I should have clarified! I just knew the OP was in the :usa: so please take no offense :D ... PS: I'm always jealous of you guys when it's winter here!!! :p:

[One_Hertz]

I don't understand why everyone says these temps are fine. These temps are exactly what I was getting on air with a similar i7 (using thermalright 120 extreme). Surely water should be good for something. I guess I'll see when I install my loop this Friday. I really do not expect much over 70C at 4.5ghz.

The fact that it shoots up very quickly means either bad contact (which you ruled out), uneven IHS, uneven waterblock, or poor IHS to core connection. If you are willing, take IHS off, sand it, clean the core, put some thermal compound between core and IHS, put IHS back on, mount your waterblock back.

[rge]

If you look at i7s in the linx 4.5 stable thread (http://www.xtremesystems.org/forums/showthread.php?t=239175), and look at difference in realtemp minimum/idle temps versus max load temps, typically at 1.4vcore/4.5ghz you will see difference of 35-45C. My chip is about 35C at 4.5ghz, 40C at 4.6, and at 4.8 around 48C difference. (you can also judge ambients by looking at min temps, usually about 13-17C higher avg min temps over ambient at that vcore/mhz).

And the temp will always jump up immediately about 30+C since TDP goes immediately from 30-40W to 200W, then get another 3-4C warmer as water heats up (example on mine from 2.5C over ambient at idle to max 5-6C over ambient at load).

A 32C jump from idle to load is average in linx thread at 4.5ghz and 1.4vcore, mine is almost exactly same, then I go up to 35C after longer load when water temp maxes. There are a few a little better by few degrees, and some hotter by 5-10C.

Intel solders the core i7 IHS with solder tim that has a thermal conductance of ~50W/M*K. The best pastes on the market will be 10X lower than that. So removing IHS is a guaranteed way to get horribly worse temps.

You an always use a better tim like indigo xtreme, and maybe get another few C better, some of my runs in that thread were using that, so I have some relatively decent temps...but dont expect miracles at 200W TDP.

@One Hertz, OP said his ambient is 23C, can you post up a linx shot while running at 4.5ghz and 1.4x vcore at 73C on air. If you can it will be the coolest running i7 I have seen.

[Waterlogged]

Nano, GTZ is a last generation block and, according to Swiftech's own testing, is beaten by the XT by 3°C. The XT however, is not currently the "best" block out there, the data we currently have points to the Supreme HF as being the "best". It's a good bit less restrictive than the GTZ which is really good for you due to running the CPX-Pro and it should allow a decent hike in flow rate. I also would suggest. . .turning up the fans. ;)


I don't understand why everyone says these temps are fine. These temps are exactly what I was getting on air with a similar i7 (using thermalright 120 extreme). Surely water should be good for something. I guess I'll see when I install my loop this Friday. I really do not expect much over 70C at 4.5ghz.

The fact that it shoots up very quickly means either bad contact (which you ruled out), uneven IHS, uneven waterblock, or poor IHS to core connection. If you are willing, take IHS off, sand it, clean the core, put some thermal compound between core and IHS, put IHS back on, mount your waterblock back.

If he goes ahead and does all that work, he might as well leave the IHS off and just go with a Storm.

[lowfat]

I don't understand why everyone says these temps are fine. These temps are exactly what I was getting on air with a similar i7 (using thermalright 120 extreme). Surely water should be good for something. I guess I'll see when I install my loop this Friday. I really do not expect much over 70C at 4.5ghz.

The fact that it shoots up very quickly means either bad contact (which you ruled out), uneven IHS, uneven waterblock, or poor IHS to core connection. If you are willing, take IHS off, sand it, clean the core, put some thermal compound between core and IHS, put IHS back on, mount your waterblock back.

But were you using 1.4V and 4.5GHz though? His temps are fine for those kind of clocks & voltage. Although I was just using a GTZ temps were ~ 79˚C for me @ 4.5GHz,1.3375V. Using 3x120 and a 2x120 radiators.

And core temps always jump up fast i7's. I think it is just a more accurate temperature sensor :shrug:

[Vinas]

I don't understand why everyone says these temps are fine. These temps are exactly what I was getting on air with a similar i7 (using thermalright 120 extreme). Surely water should be good for something. I guess I'll see when I install my loop this Friday. I really do not expect much over 70C at 4.5ghz.

The fact that it shoots up very quickly means either bad contact (which you ruled out), uneven IHS, uneven waterblock, or poor IHS to core connection. If you are willing, take IHS off, sand it, clean the core, put some thermal compound between core and IHS, put IHS back on, mount your waterblock back.One Hertz, I think you are absolutely correct in that you should be getting better temperatures than you did with your TRUE. In the case of the OP, he seems like he has a pretty warm ambient temperature coupled with a high vcore, big OC, and HT enabled. In my experience i7 vcore much above 1.38v and 4GHz leads to really blistering hot temperatures in an instant under linpack.

[One_Hertz]

His HT is OFF. I think some people are forgetting this... The temps WOULD be fine if HT was ON. He gets to 72C immediately before the water even has a chance to heat up. rge - in your very link everyone is running HT ON. There is like a 5-10C difference. On top of that, even with HT ON almost nobody is hitting 40C idle.

And yes I was getting ~75C with 4.5ghz (or just under that) with 1.41 volts with HT off using thermalright extreme (~2500rpm fan) with 20C ambient on an i7 860. I can't screenshot because my current load temp is -15C with phase. That was just a quick test I ran before mounting phase on.

[Vinas]

One Hertz: My bad about the HT, I was thinking of a different thread. :shakes: Hopefully when you switch from phase to water, you'll be satisfied with the temperatures. Just because I think his temperatures are pretty normal for his speed range doesn't mean I think we should be okay with it. :D Lets find ways to help the OP lower his temps! :up:

Would you really risk popping the core by taking off the IHS? Would TIM transfer heat better than the silver solder that intel uses? I'm a little skeptical that its worth the risk in removing the IHS, but would agree that a bad mount and/or uneven IHS could cause poor temperatures.

@ OP: are you pretty confident with your TIM application? Care to share what kind and how you applied the TIM?

[Vapor]

It could be a few things...though the temps shooting up instantly is a very normal thing (then they climb slowly as water temperature reaches equilibrium, in this case from 72C to 77C). 40C at idle is pretty normal but 77C with HT off is a bit high for these settings.

Anyway, here's my thoughts for what it could be:

1) It's just a hot CPU. Nothing you can do about this other than get a new CPU if you don't like it (though it is my opinion that if you have clocks you want and the stability you want, that's all you need).

2) TIM is bad. What TIM are you using, Nanometer?

3) Block inlet/outlet are reversed.

4) Block orientation is off by 90 degrees.


Your temps aren't that far off what I would expect for a median chip and if you eliminate all the possible things you have influence on, then you're left with only a couple options:

1) Your chip is hot.

2) Improve the CPU block.

[One_Hertz]

Would you really risk popping the core by taking off the IHS? Would TIM transfer heat better than the silver solder that intel uses? I'm a little skeptical that its worth the risk in removing the IHS, but would agree that a bad mount and/or uneven IHS could cause poor temperatures.

No, the solder Intel uses is much better as mentioned before in this thread, but the problem is, sometimes they :banana::banana::banana::banana: up and it is not making proper contact so it is not just about the solder. Of course this is rare but it is a possibility.

OP should have around 5C less IMO.

[rge]

His HT is OFF. I think some people are forgetting this... The temps WOULD be fine if HT was ON. He gets to 72C immediately before the water even has a chance to heat up. rge - in your very link everyone is running HT ON. There is like a 5-10C difference. On top of that, even with HT ON almost nobody is hitting 40C idle.


Your right, I missed his HT being off (so your supposed to read all the posts now before posting:p:), hence I assumed it was on. And agree, definitely useful troubleshooting things previously mentioned with tim/mount/wb, etc.

My current ambients are 25.6C (2.6C higher than nanometer's).
At 4.5ghz HT off, 1.41vcore, 1.38 vtt...

idle temps, 41-42 with HT on or off, maybe 1C higher with HT on. (I have gtx 295 in my cpu loop, so idle temps may be little higher than most).

Instant prime load (30 secs prime) I am way cooler, ~66C avg and that is with 2.6C higher ambients. (35+W difference at load HT on/off.

So yep I agree, either his cpu is hotter than mine, or some room for improvement exists.

[One_Hertz]

Surely water should be good for something. I guess I'll see when I install my loop this Friday. I really do not expect much over 70C at 4.5ghz.

My temps with 4.5ghz, 1.44 volts, 1.39vtt, ht off, 20c ambient are 30C idle, 62C prime load (starts at 59C then water heats up a few degrees).

[Cookiesowns]

I run my 920 3836A761 C0 with 1.49 volts no HT at 4.2 and it barley hits 90c with a Hyper 212+ after 5-10 minutes of PRIME95 ambient of 27c? So I'm assuming either you have a defective WB, or just your CPU is amazingly Hot..