2500K Temperatures

[anasuelli]

Hi everybody,

I just got my 2500K with Asus Sabertooth P67 Rev3. I installed my good-old Ultra 120 with a pair of Noctua FLX fan and silver thermal compound. Until now I tried 4.0 GHz and 4.2 GHz without any Vcore changes; seems stable but I'd like to compare my temps readings with yours; notice that I took my reading with an ambient temperature about 27 C.

Idle (Windows 7 Desktop): 35-38 C
Load (Metro 2033 Benchmark): 50-55 C
Stress (max over Lynx 64bit): 79-82 C


Those seem a little too high for a Sandy with TRUE dissy... isn't it?

I cross checked the temperatures with CoreTemp, OpenHardwareMonitor and OCCT4Beta.

Thank you in advance.

 

[micropage7]

it looks hot but compared to my stock speed 3,2 gig when i run linpack it can hit about 60 degree so around 80 on 4gig looks accepted but you need to pay attention to it
try to improve the air flow, if you wanna try lap the heatsink or consider water cooling

 

[Crap Daddy]

It seems a bit hot. Not that it can't take it. Something might be wrong with HSF mount. I have a cheap and reliable Hyper TX3, just as I'm typing I have 28-30C. Under prime, 4.2 just by upping the multi I had mid 60C with fan speed set to increase after 60C.

 

[FreedomEclipse]

Even when i had my Frio, running with 2 Akasa Apache's my max temps never went beyond 71'c

with my Antec 920, it never goes above 60'c

 

[micropage7]

try to re seat the hsf again and reapply the paste again maybe the heat transfered from processor to hsf not good so it could from your hsf not tight, lack or too much paste, bad hsf contact, or the base material or the fan cant remove the heat completly

 

[anasuelli]

Thank you guys for your quick replies!

Unfortunately I already tried a re-seat of the TRUE and the results didn't change.

There are two things I didn't tell you: the heasink base is not lapped at all, it's very old and I used it on multiple processors over the years, maybe the base lost a lot of performance over time.

Also my new fans (Noctua FLX series 12mm) have a different behaviour then the ones I used before: they blow more air with less noise but I suppose the static pressure is lower, because I can feel a lot of air bouncing back instead of passing through the heatsink fins.

So my questions are:

- Can a "ruined" base (not in the sense of destroyed, lol, it doesn't have visible scrathes) cause such high temps?
- This very same heatsink managed my previous cpu very well: an overclocked and heavily overvolted Q6600 G0! So what? Are those temps completely off in your opinion?

An interesting thing is that the behaviour between the two processors differs a lot: my old q6600 was obviously hotter in idle (little less then 50 C on the desktop), pretty similar in load (about 60 C) and incredibly cooler then my actualy 2500K during Lynx (never saw above 70 C with the Q6600).

In other words, this damn 2500K is freezing cool in idle, acceptable in load, but with Lynx it goes 70 C in a pair of seconds and goes well above 80 after a minute!!

I just wanted to ensure that this behaviour is normal.

I really have to thank you ALL for your help. You are cool (while my cpu isn't)

 

[FreedomEclipse]

What thermal paste do you use??? maybe it has a long burn-in period??

 

[anasuelli]

What thermal paste do you use??? maybe it has a long burn-in period??

It is possible! How much does it takes usually for the compound to reach it's best chemical state?

I am using a cheap 30% silver Nexus compound. My arctic silver tube was deplete so I choose to use that Nexus. Do you think it's been a bad idea?

 

[Millennium]

anasuelli,

I think those temps are normal for this cooler. I have the same one with 2 (quietish) aftermarket fans and my 2500k @ 4.5 hits 85 degrees with it :/

Still thats intel burn test i dont think anything else sends temps so high so I'm ok for now.

edit: I used ceramique

 

[FreedomEclipse]

It is possible! How much does it takes usually for the compound to reach it's best chemical state?

I am using a cheap 30% silver Nexus compound. My arctic silver tube was deplete so I choose to use that Nexus. Do you think it's been a bad idea?

That thermal compound seems similar to AS5. so im guessing it has a burn in period also.

Id just let the system keep running and check the temps again after 8-12hrs.

I personally use AC MX-4. easy to spread, and theres no burn in period.

 

[anasuelli]

I hope Millenium is right, but take a look at this:

http://www.hwupgrade.it/forum/showthread.php?t=2379254

it's the main italian hardware forum, 2500k/2600k oc thread. The second post lists a bunch of user's tests.

As you can see with i5s and voltages between 1.168 and 1.272 V and frequencies between 4.2 and 4.4 GHz, and similar aftermarket heatsinks, temps with Lynx are much lower then mine. And those written there are the max temps reached in a 10-cycle lynx run. My temps are about 15-20 degrees higher with just 4 or 5 cycles.

That is what makes me worried.

 

[Millennium]

I have to run 1.38 vcore to get 4.5ghz on my chip!

 

[FreedomEclipse]

I have to run 1.38 vcore to get 4.5ghz on my chip!

I can do 4.8Ghz@1.36

Youve either got a really shit bad chip or your motherboard is shit bad.

 

[anasuelli]

I'm 4.4 GHz now, still no manual changes on Vcore, but I let the Load-Line Calibration on Auto so the real voltage in full load floats between 1.28 and 1.30 V.

I hit 91 C on Lynx.

Again: in real-use full load (playing games, using heavy image/video editing/transcoding) it doesn't get over 55 C even after hours of continuous load.
Lynx send it to 90 degrees in a bunch on minutes. I know this is normal but.... 90 C? seems crazy to me.

Now the burn-in period should be ended.

Can anyone else with this processor and a similar heatsink post their max temps over Lynx 64 bit (5-10 minutes).

Thank you in advance, again.

 

[arnoo1]

I have the same hs as you and my 2600k never goes over 60c @ 4,3ghz 1.3v bios, i have push/pull noctua's, maybe reseat your hs or you have a "hot die"

 

[anasuelli]

I have the same hs as you and my 2600k never goes over 60c @ 4,3ghz 1.3v bios, i have push/pull noctua's, maybe reseat your hs or you have a "hot die"

Sorry to bother, I appreciate your reply very much, but I just want to be sure we are talking the same thing: when you say "my 2600k never goes over 60c" what do you mean precisely? You mean that in your daily use "full load" it never goes over 60 C? If that so, my 2500K too... it never goes over 55 C actually neither after hours of intense 3D games.

Different thing if you are talking about Lynx: if your phrase "my 2600k never goes over 60c" is intended to be ""my 2600k never goes over 60c IN LYNX", I will start to worry.

Please respond!

 

[arnoo1]

Sorry to bother, I appreciate your reply very much, but I just want to be sure we are talking the same thing: when you say "my 2600k never goes over 60c" what do you mean precisely? You mean that in your daily use "full load" it never goes over 60 C? If that so, my 2500K too... it never goes over 55 C actually neither after hours of intense 3D games.

Different thing if you are talking about Lynx: if your phrase "my 2600k never goes over 60c" is intended to be ""my 2600k never goes over 60c IN LYNX", I will start to worry.

Please respond!

If i run prime95 or linpacks for ours, it never goes over 60c average

 

[anasuelli]

If i run prime95 or linpacks for ours, it never goes over 60c average

Thank you for your feedback, I suppose something's wrong here. I'll keep searching for the true reason for this heat-up.

 

[FreedomEclipse]

Check if the heatsink is making good contact with the CPU. use a credit card to spread the paste. maybe the heatshield on the CPU or the base of the heatsink might not be entirely level.

again im drawing at straws here. maybe you were just unlucky enough to pick up a hot CPU

 

[anasuelli]

Check if the heatsink is making good contact with the CPU. use a credit card to spread the paste. maybe the heatshield on the CPU or the base of the heatsink might not be entirely level.

again im drawing at straws here. maybe you were just unlucky enough to pick up a hot CPU

I'll keep you posted about this, in the meantime, you all get an endorsement for your collaboration.

Again, thanks.

 

[johnnyfiive]

Hi everybody,

I just got my 2500K with Asus Sabertooth P67 Rev3. I installed my good-old Ultra 120 with a pair of Noctua FLX fan and silver thermal compound. Until now I tried 4.0 GHz and 4.2 GHz without any Vcore changes; ....

Thats your problem right there anasuelli. It's probably using 1.37v+ if you're leaving it on auto voltage settings, which it sounds like you are.
For 4.0-4.2, you only need 1.25v (varies but thats typical) for 4.2. At 1.25v, with that cooling setup, you should be hitting around 55-65c MAX with a 27c ambient.

 

[xBruce88x]

just out of curiosity, what case and fan setup do you have for this rig? (aside from the cpu cooler)

 

[arnoo1]

my 2600k running there for a hour

 

[johnnyfiive]

^^^ Exactly, those are normal temps for 4.2. Hopefully the OP got his heat issue worked out.

 

[Spaceman Spiff]

Aaah these threads drive me nuts!

OP can you post your solution? It may help others and I am curious!!

That's why my friends call me whiskers....