I7-950 Temperature Values

[Timon5022]

Hi all, I've been roaming through some of your posts lately due to the fact that I've just bought myself a new PC and as a CPU I chose the i7-950. With it I bought the aftermarket cooler 212+ from Coolermaster to make sure that I could OC the CPU and keep the temps decent.

However, I feel like my temperatures are really high and I saw some of you discussing what the highest temperature should be. Some say that 68*c is what I should aim at, 73*c being the maximum, While some argue that it has to be under the tjunction of 100*c. As you can see, that's a lot of difference, and I'm not at all used to reading temperatures close to 100*C from my old E6750.

At the moment I have it clocked at 3.6ghz and when running the intelBurn test it reaches temps around 73 (the highest I've had for a brief moment after sequential tests on very high settings). Also I have HT disabled to keep these temperatures.

So my question is two-fold. What is the ideal temperature to shoot for if I want to get the most out of my OC, without being in the risk of shortning the life of my CPU noticeably? And are my temps too high? Have I done something wrong mounting the CPU Cooler? I have tried reseating it, but it still idles around 40*c.

Ofc I don't know what the ideal temps are, as I'm not used to this CPU, but I am a bit "disappointed", since I went out of my way to get a real nice Case (the corsair 600t) and made sure all the cables were routed behind the motherboard so it should have surpreme airflow. But maybe this is what I should expect? And ofc, if the "safe" temperature goes all the way up to under the tjunction then I still have a lot of headroom But that's for you to tell me

Thanks in advance

[burebista]

Keep your distance to TJmax>20 in daily use and forget about temps.
Also Linpack clones put an unrealistic load so you'll never see those temps in daily use.
Those temps are core temps measured by RealTemp/CoreTemp?
Ambient temperature and Vcore plays a major role too in your core temperatures.

[PatRaceTin]

212+ May not able to hold
overclocked 1366 CPUs

[Timon5022]

I have set my vcore at 1.25 and I'm measuring the temps with RealTemp, have used Coretemp and others but now I use RealTemp as I've read on this forum it should be more accurate. Also ambient temps should be fine I mean they are probably around 20*C, it's not like it's super hot in here and as I said the case airflow should be very good.

212+ May not able to hold
overclocked 1366 CPUs

Ah, what cooler do you suggest then? I've been looking at the H70 from corsair Looks nice and good and easy to deal with.

[Cursed_Wings]

the H70 is basically a waste when the H50 will show the same performance with optimal fan setup. as long as you're running under 80C during stress tests you should be fine. if you really want to upgrade your cooler the cogage true spirit is a good cheap alternative for under $40, it shows the same performance as the current top 120mm single tower coolers. taking it to the next level with air cooling would be moving to the dual tower coolers such as the thermalright silver arrow/cogage arrow or the nh-d14, take note that they're quite large though.

[Leeghoofd]

You got data to back that claim up Cursed wings regarding the H70 vs H50 ?

On I7 950 I get almost 7°C difference between the H50 with dual fans at 1600rpm and the H70 with it's fans running at 1600rpm... (4.2Ghz at 1.4Vcore) With the i970 Gulftown the difference is even bigger (also 4.2Ghz at 1.42Vcore). The H70 is able to hold the load far better than it's little brother...the latter even with a dual fan setup... I couldn't keep the 1090T stable with the H50, with the H70 I can...

[Cursed_Wings]

testing done by SonDa5 on hardocp forums here; http://hardforum.com/showthread.php?t=1555079

tests were done using 38mm fans with shrouds in push-pull.

[Leeghoofd]

Tim spread out on H70 is very thick, esp near one side (which to me is not a perfect mount then). Little Vcore discrepancy 1.36 vs 1.38 (although nothing was changed in the bios...)

Why more fans ? It isn't always better, also the same for higher RPM ones (noise vs cooling efficiency can be totally crap)... Some rads are designed to be used otherwise... it's not because it worked with the H50 that it also is better with the H70...

Doing stock H50 with single and dual fan setups and the H70 with the fans at blast vs low noise ( 0.5-1°C there difference max), H70 knocks the socks off the H50 when being used with a HOT CPU (read 6cores,high frequency, High Vcore...setups) In all my tests S775,S1156, AM3 and I7, I got better results with the H70. Especially with the 970 the difference was mindblowing...

Like I said before I couldn't pass the tests on the 1090T with the H50 (single fan) as easy as with the H70 (low rpm)... At a certain moment I was tempted to redo the tests at a lower clock speed. Regarding stability especially AMD is very temp sensitive.. a clear indication to me that the H70, out of the box, is superior when it has to cool down a heavy loader...(again) talking all about stock setups here (with the discrepancy of adding one fan)

Let him check the pump speeds too... something is not 100% right. H70 should be better then H50, for sure when used with the stock fans...

Last call here : if you add all of them fans, you loose the easy mounting in any case and all the fans add up to the extra cost... which completely ruins the idea of it's concept and other easy to mount solutions alike the CoolIT ones...

To sum itup : An H50 with dual fan setup is a must (single fan it doesn't cut it anymore with modern CPU's), an H70 with the original fans at +/- 1600rpm has got similar or better cooling capacities then the dual fan H50 setup (also with Corsairs fans spinning at +/-1600rpm)... Running the fans at high speed (2000rpm) is not a good idea for your roommates !!

[Cursed_Wings]

a single san ace 9G1212H1011 with shroud would probably show similar results to SonDa5's testing seeing at the scythe kaze are weaker in comparison, they're also very quiet for 38mm fans (39 DB manufacturer spec. which should be close to true levels seeing as it's not really a consumer fan). they can also be found on performance pc for $23, so total cost of the H50+fan would still be under the cost of the H70 (unless it's found on sale) and size wouldn't be an issue compared to the H70 stock push-pull.

all in all, the H70 will perform better than the H50 with weaker fans or at lower(silent) rpm, but with an more optimal setup I still have to say it's not worth the cost.

[Timon5022]

Hi again, I've got a stable OC at 4.06 Ghz atm, and when running Intel BurnTest my temp reaches 82, 18*c from tmax... Would it be sensible for me to push it even more? or would the daily temps, when running games and such reach too high temps?