CPU Temperature

• Invertex wrote on 2012-01-08 08:19
  Well, my CPU fan broke on me earlier today. My dad helped me replace the fan and reapplied the thermal paste (I think that's what it's called) to it.
  
  He's making me run some stress tests on the CPU to test the temperature.
  
  I've been running it for 25 minutes now. The max temps are 90, 90, 86, 86 (Degrees celcius, for each core). I'm not sure what's the safe range I should be aiming for (I think it was <80 under load?), and I'm not sure if I'm missing anything that'd make my system a bit more stable.
  
  So I guess it's not much of a question, but any suggestions?
  
  The CPU is an Intel Core 2 Quad Q6700.
• Senbonzakura wrote on 2012-01-08 08:36
  Those temperatures sound unusually high. How long has it been since you cleared out dust from inside your rig?
  
  EDIT: I'm also curious on how the thermal paste was applied. :P
• RebeccaBlack wrote on 2012-01-08 08:36
  It depends who you ask. CPUs tend to have loooong lifespans, so even if the lifespan is reduced, the part would often become obsolete before anything bad would ever happen to it. Others will tell you this is a bad idea and they like to keep it under 80 to be safe. I'd wait for more opinions on this, but that's my general understanding of it. Definitely be careful though, if it starts to go up over 90... it could be a very serious problem.
• Mystickskye wrote on 2012-01-08 10:10
  How are you testing it? Those temps are high though not alarmingly (When I say alarmingly here I mean like, house is on fire alarmingly) so. If that's as high as it goes after 25 minutes of pure stress testing then you should be okay.
  
  I'm assuming it wasn't this high before though, is the fan different or anything?
• Invertex wrote on 2012-01-08 10:34
  Well, it's a fairly new computer. There isn't much dust on anything yet.
  The thermal paste was applied directly the the chip, and then the fan was put on top of it.
  
  And it was after about 25 minutes of 100% load on all cores.
  
  I finished the test with a total test time of one hour. Maximums of 91, 91, 89, 89. It stopped getting any hotter afterwards, and bounced between 86 and 90 for the first two cores and 85 and 89 for the last two.
  
  I don't think it should be going for over half a hour with all the cores at 100% though. ;-;
  
  Idle temperatures seem to range from 30 to 45.
• Mystickskye wrote on 2012-01-08 12:38
  Well, it certainly doesn't sound right. That's how hot my 2500k would get if I OCed it over 4GHz on the stock cooler :| I was thinking old computer and all but if it's new...
  
  Are you able to get specs/photos of the various parts involved? Also do you know what the temps were like prior to the change?
• Invertex wrote on 2012-01-08 15:06
  Well, I just asked my dad about it.
  
  It's apparently a relatively old fan, and he has a better one. He just wants to see if this fan (had an identical one that was working fine) works as the other one is probably a pain to install.
  
  Errr, I don't know the specifics of the build. I don't know the name of the motherboard.
  I know the graphics card is a GTX470 and it has 6GB RAM. The box for the PSU says it's an OCZ Evostream or something. The CPU itself isn't overclocked.
  
  Dad also told me that prior to changing the fan (which occured at the same time as changing the graphics card and PSU), the tenperature during stress tests were perfectly fine.
• Yoorah wrote on 2012-01-08 19:37
  Those are a bit high. But keep in mind that typical CPU demanding tasks won't generate as much load as a stress test. And Intel's first quad cores were known to get hot. Do not compare them to the 2500K, as they are built on a different process, have different TDPs, etc.
  
  My guess is that the thermal paste was not applied as well as it could have been and/or the heatsink wasn't installed securely.
• TA wrote on 2012-01-08 20:10
  Those temps are way too high. Most CPU's will melt around those temps.
  
  60 under load would be what you're shooting for iirc.
• Yoorah wrote on 2012-01-08 21:09
  60-70C under real-world load, yeah. But it's normal for a stress test to go higher, especially if it's on stock cooling. And you gotta remember that back when Core 2 Quad came out, many were complaining about how hot those chips got. Still, 90C is approaching shutdown temp... so yeah.
  
  Take the heatsink off and look at the footprint in the thermal paste. I'm guessing there's an air gap. Or the heatsink wasn't secured fully (I hate Intel's plastic clip heatsink mount garbage).
• Invertex wrote on 2012-01-08 21:58
  Well, guess I'll be checking the fan later.
  Potentially not until next weekend though, since I have school tomorrow. ;-;
• Mystickskye wrote on 2012-01-09 02:28

  Quote from Yoorah;728006:

  And Intel's first quad cores were known to get hot. Do not compare them to the 2500K, as they are built on a different process, have different TDPs, etc.

  Sorry, my point was that such temps were out of the ordinary (even the for "hot" early models). Should have used a better example, my bad! Looking around for more relevant info, others running stress tests with the same CPU don't get such high temps either unless overclocking and even then not as high as Invertex's (admittedly, these examples I'm seeing aren't massive OCs).
  
  But yeah, at this point it just seems that something's up with the heatsink. My guess is what yoorah said, insecure installation.
• Yoorah wrote on 2012-01-09 02:54
  Yeah. I tried to look up numbers before, but results had people using various aftermarket coolers, including water cooling. lol I'm guessing Invertex is using the stock HSF. I doubt the stock HSF handles a Q6700 undergoing a stress test too well. xD;