• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

i7 920 load temps...are they normal?

computermod14

computermod14

Limp Gawd
Joined
Nov 6, 2003
Messages
327
I currently have my i7 920 @ 1.325V (D0) and it idles ~31C with an ambient room temp of 77F. Under load (prime95 particularly) it may reach 75-77C. All specs are provided in my sig and as you can see I am using a triple rad with a push/pull setup. I'm wondering if anyone out there is running a similar setup that can compare to my scenario. Thanks! :cool:
 
I have a 920 C0/C1 at a little less than 1.3V at 4GHz. I idle between 34-40C, but I only reach 58-62 at full load.

I have the 360 GTX and a 120 GTS, but I am currently not using push/pull. I also have both my two 6950s in the loop, so to me your load temperature seems quite high, but I don't know how you manage so low idle temps.
 
I think your load temps are high too. Yours are like mine were and my block install was fucked. Now at 4.1 I am at 63c with prime95 torture test. Same ambient as well.
 
What would you consider I try doing? I know the fans I use are not the best for static pressure, I mostly got them for the UV blades for my color scheme. They are not the worst fans out there either... Also, the idle temp I gave you is not any of the core temps...core temps range from 39-45C idle and 80-90C load (prime95). I have already reapplied my AS5 following the instructions from their site which is the straight line technique. This is not my first build either...

33cbtoj.jpg


6fy420.jpg
 
Last edited:
someitmes it takes several reseats before you are satisfied. all overclocking results with vary maybe you just ahve a high leakage chip. you can try lapping it but that is an experience on its own and some people are wary about toying with a cpu like that. i did it on my 930 and decreased temps about 5c not a huge imporovement but some non the less.
 
With a GTS 420 and 3 140mm 1500RPM fans in a CPU loop only, my 920 peaked out around 70 C with 1.281v at 4.2 ghz. Idle was around 35 C.
 
I am in a similar situation.

I use a Scythe Mugen 3 as cooler, idle temps are around 37 C, under 100% load it goes up to 78 C. I used to have a Scythe Ninja 2, which had the same problem. Switching to the Mugen 3 did not help. Reseating it and tightening the screws did not help. And my i7 920 runs at factory speed (2.67 GHz).

The odd things is, when touching the heatsink under heavy load it feels almost cold, not like 78 C. On the other hand, as soon as I stop stressing the CPU, the temperature drops to around 40 C almost immediately. So on the one hand, the cooler does not seem to dissipate the heat, on the hand it seems to do it well enough.

I am on the verge of ordering a Corsair Hydro Series H100 Extreme or a Antec H2O 920 to see if that will help get my temps down. What do you guys think?
 
You have really poor contact. You're doing something wrong if your processor is at 78 C and your heatsink doesn't feel warm.

Even with my watercooling setup, I can feel the heat coming off when the processor is under load.
 
Tsumi said:
You have really poor contact. You're doing something wrong if your processor is at 78 C and your heatsink doesn't feel warm.

Even with my watercooling setup, I can feel the heat coming off when the processor is under load.
Click to expand...

I was thinking the same at first, but if that is really the case, I would not expect the temperature to drop almost 40 degrees as soon as the load stops. But in any case, I guess I will try to reseat that cooler one more time...
 
quadrox said:
I was thinking the same at first, but if that is really the case, I would not expect the temperature to drop almost 40 degrees as soon as the load stops. But in any case, I guess I will try to reseat that cooler one more time...
Click to expand...

That actually is completely normal, and expected. Mine shows the exact same behavior; as soon as the load comes off, it drops ~35-40 C. The CPU does not store heat very well, and as soon as the load is gone, the heat is dissipated instantly.
 
Seing that your voltage are pretty high and depending on ambient temperature / humid. Give or take Loops seem to be within range. Also if you have gpu in the loops it does add to the heat, 4-10c per card.
 
Well this is the proactive step that I've taken to try to improve the temps...

3x120mm NZXT FX-120LB for push config with 3 20mm Phobya shrouds in between
3x140MM NZXT FX-140LB for case fans
Switching to 1/2ID tubing

Lets see how this pans out for me once I receive the items in the next few days. I will be back to update for yall.
 
Tsumi said:
That actually is completely normal, and expected. Mine shows the exact same behavior; as soon as the load comes off, it drops ~35-40 C. The CPU does not store heat very well, and as soon as the load is gone, the heat is dissipated instantly.
Click to expand...

Ok, that's good to know. I just reseated that damn thing again, and it still shows the exact same problems. CPU was displayed as 80C, but even the baseplate directly above the CPU was not warm.

Could this be a sensor readout error? I just find it extremely hard to believe that I am completely unable to get either of my heatsinks mounted properly. powertop2 on linux doesn't even show that the CPU clock goes down, shouldn't it do that completely on it's own when it gets hot?

this is just so damn weird.

EDIT: Same behavior on windows with prime95. Probably not a cpu readout problem then. Why won't the damn heatsink work?
 
Last edited:
The reason for the large drop is because of power gating. As soon as the cpu isn't using a certain part of itself, it gets shut off. And as mentioned, the heat dissipates almost instantly(goes from making heat to making no heat, but your cooler keeps dissipating at the same rate).

Anyway, your temps are not dangerous where they stand. Your CPU will last well beyond its useful lifespan. There could be a number of explanations for high temps tho - poor contact(warped motherboard, not enough tim, uneven mounting), not enough TIM, air in your TIM, or you could just have a leaky cpu. Temps can vary wildly from chip to chip. Literally, two cpu's running exactly the same settings with exactly the same coolers can have up to a 15C difference due to differences in gate leakage.
See:
vs
The left is my current chip, which runs 4.6GHz HT on, with 1.525v BIOS. Maxes out at 84C.
The right is my old chip, which runs 4.5GHz HT on, with 1.525v BIOS. Maxes out at 94C.

Same exact voltage, but the chip running faster is COOLER... Significantly!!!

If you want to really stress test your cpu, get LinX, and update to the latest Intel MKL. Don't forget to boot into diagnostic mode and use 'all' memory for the test.

Linpack puts Prime95 to SHAME. Expect your cpu to break 85C, maybe even 90C with your exact config.

Diagnostic Mode:
Start -> Run -> msconfig -> click the diagnostic mode button -> ok -> restart
 
Last edited:
Ok, I have switched back to the Ninja 2, it seems to cool slightly better than the Mugen 3. (idle temp around 33 C, instead of 37 C). Oh well, I'll just have to live with the high temps, I suppose. It doesn't affect performance, and if it eventually burns out I will just have to buy a new one.
 
Load temps are way to hight. Ive got a 940 co that gets those temps at 1.41 vcore with a 360 rad and high speed push pull fans.
 
For giggles I figured I would return all BIOS settings to default and rerun temp scanning. This is what I've come up with...seems like a pretty drastic change, of course it's jumping from 1V to 1.325 though

2wpifcx.jpg


vnnmsx.jpg
 
that heatsink is cocked, you are running 20C difference across the chip
 
one side of the chip is cooling better than the other, I actually had this effect recently on an old dual core athalon a buddy needed the heat sink modified to fit a video card...long story, anyway when I first lit it back up one core was way hot...sure enough I had not got the sink flat on the chip

I think you are seeing the same thing, good contact on one edge or corner but the other side is not
 
Whenever I mount my waterblock (remounted a few times) you can only turn the thumb screws so far down until they stop. I turn each one about 3/4 turn at a time all the way to the end. The only other thing I can think to check is to take a ruler to the top of the CPU to see if there are any dips
 
could be a lot of things, little grain of sand got in there with the TIM, warped backplate, bad thread on one screw

I dunno, hate to tell you to tear it back down. If you do, try to mentally step back from it and think, examine all the parts carefully looking for an explanation.
 
Ok, I got myself a Corsair Hydro Series H100, and it has helped quite a lot. I can keep the temp below 58 C now, even at 100% load. That's good enough for now.

A large part of the improvement probably comes from the fact that it is so easy to install with 4 thumbscrews and plenty of space around, so that it is much easier to make it have good contact with the CPU.
 
Yeah, I've burned myself touching the heatsink while CPU under load. But those load temps are what I'm getting on air. Those 920s are hot chips.
 
You're using the EK supreme on the cpu right? Are you using the backplate? I have the CU version of the same block and had to leave the backplate off as it sat crooked on my mobo (sig). It actually made the mounting bolts sit at about a 3 degree angle.
 
I have the EK Supreme block on a GTX 360 radiator with 3 Gentle Typhoon 1850 rpm fans in a push config (pushing out).

Running 4.4GHz, HT on, 1.393v @ 210bclk I hit mid/upper 80s. Idle is around 35-40c depending on the core.

I am switching to a push/pull with 3 more GT1850s, pulling in and swapping the rear fan on my 800D case to exhaust. Hoping this will drop temperatures down. I definitely did not have my block seated correctly the first time, I reseated it and all four bolts are as tight as I can physically get them on the motherboard, so I don't think my issue is poor contact. Your temperatures are pretty much in line with mine, although I have more voltage and MHz.

I don't know how some of you are getting load temps in the 50s at 4+ GHz. You must have some very efficient chips, or mine must be very leaky. In any case my CPU has been running this speed as my daily 24/7 PC for a year with no issues so I am not super worried.
 
So after reading this thread I decided to redo some of my case cooling + radiator setup. Switched my fans to pull in instead of push out, and changed the rear 140mm fan to exhaust.

Then I remounted my waterblock and right away I caught something I missed the first time. I did not use the plastic spacers included with my EK Supreme HF, which was causing uneven contact. Unfortunately I can only find 2 of them :( But I installed 2 of the 4 and was much more careful. I am now seeing temps in the high 70s load and low 30s idle @ 4.4GHz, with core1 hitting 83C load (the rest hover between 76-79).

MUCH better.

I need to see if I can find a set of those spacers..
 
Well, I deffinitely have the backplate on and am using all 4 spacers. I did not happen to see any slight angle of the four mounting bolts but I will be tearing my whole system down tomorrow to prepare for my new watercooling equipment. I also lapped the chip tonight and reseated the waterblock on the chip. I reran prime95 and temps seem to be about the same as I originally posted, with that 3 cores running warmer than the 4th.
 
Well folks, this is what I ended up with after lapping the CPU, reinstalling the backplate, upgrading to 1/2" tubing, and changing the fans to NZXT FX120LB's.

2vj1roz.jpg


2rftssw.jpg


And of course how it looks!

23usym1.jpg


Little improvement, guess a lot of it just has to do with the chip. The voltage has also been decreased from the original screenshots but I beleive temps would still be lower than originally. Thanks for the help guys...
 
Looks good. Those temperatures are fine too, nothing to worry about.
 
Maybe the heatspreader of your CPU and/or cooler surface isn't as flat as you think. I often read people talking about the thermal grease being useful for filling "microscopic" gaps between the CPU and cooler. Microscopic? I took a knife (on the blunt end) and a paper razor (again on blunt end) and put them against the CPU and cooler surface and then look at them against a bulb light. Microscopic my a$$. My Scythe Shuriken would let see the light of the buld on most of its surface. My Deepcool Edge400XT was far worse, showing very poor flatness (i could see the light passing everwhere except from the 2 outer edges and on both axis. My Rasetsu is very flat for like 2/3 of its surface and then is slightly bent upwards on the 1/3 of one outer side. The CPU itself seems to have a "bump" on the heatspreader that makes contact before every other point of the heatspreader does. For that i am 100% sure also by verifying it after 4-5 re-application of thermal paste. That "bump", when removing the heatsink was always almost paste-free no matter if i used the pea method or the "spread" method.
 
23usym1.jpg


LOL,Your fan are using case air to cool your water, you would get better temps if you turn the fans around and use the ambient air on top of the case.
 
spitter said:
LOL,Your fan are using case air to cool your water, you would get better temps if you turn the fans around and use the ambient air on top of the case.
Click to expand...

Not necessarily, I have a 140 below the cam shot which is pushing a lot of ambient air upwards...

Faethon said:
Maybe the heatspreader of your CPU and/or cooler surface isn't as flat as you think. I often read people talking about the thermal grease being useful for filling "microscopic" gaps between the CPU and cooler. Microscopic? I took a knife (on the blunt end) and a paper razor (again on blunt end) and put them against the CPU and cooler surface and then look at them against a bulb light. Microscopic my a$$. My Scythe Shuriken would let see the light of the buld on most of its surface. My Deepcool Edge400XT was far worse, showing very poor flatness (i could see the light passing everwhere except from the 2 outer edges and on both axis. My Rasetsu is very flat for like 2/3 of its surface and then is slightly bent upwards on the 1/3 of one outer side. The CPU itself seems to have a "bump" on the heatspreader that makes contact before every other point of the heatspreader does. For that i am 100% sure also by verifying it after 4-5 re-application of thermal paste. That "bump", when removing the heatsink was always almost paste-free no matter if i used the pea method or the "spread" method.
Click to expand...

I don't beleive the it's the waterblock...it is nickel finishing and it's so flat that it could be used as a mirror when clean. This is the reason I didn't try lapping and will not lap the waterblock.
 
spitter said:
LOL,Your fan are using case air to cool your water, you would get better temps if you turn the fans around and use the ambient air on top of the case.
Click to expand...

Shows you how much a newbie knows about watercooling :rolleyes:
 
looks nice man...

that 10C variation across the chip is odd, but maybe it is just how your specimen is. I am satisfied at this point that your mount is good.
 
You must log in or register to reply here.
Back
Top