The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.
Scary high CPU temperature [Solved]
Salvius
Registered User regular
Registered User regular
I have an E8400 (technically Xeon 3110, but same thing) with the stock Intel cooler. Not overclocked or anything, of course. It's on the Gigabyte P35-DS3L motherboard and in a Coolermaster Centurion 5 case, if it matters. I also run Folding@Home, a distributed computing program that maxes out CPU usage, though only on one core since the SMP version isn't recommended for less than quad-core.
Earlier today, I downloaded Real Temp 2.6 on a whim and it told me my CPU's core temperature for both cores was hovering around 95°C (!!). Distance to TJMax varied from 1 to 0 (As the documentation explains, TJMax is the maximum safe operating temperature; at which the processor automatically throttles down clock speed) I turned off Folding@Home and it immediately declined, finally stabilizing at around 45-50°C. When I turned F@H back on it climbed up to the same as it was originally in a couple of minutes and stayed there until I quit the program. I downloaded Core Temp, which reports the same except about 10 higher. Temperature is actually reported by the sensor as distance from TJMax, so the discrepancy is just because Real Temp thinks it's 95°C and Core Temp that it's 105°C. For both, however, the processor shot right up to the limit when put under load with F@H. Obviously, something is very wrong.
The question is, what's wrong and what should I do? I checked and the fans are still spinning fine. Is the sensor wrong? Is the heatsink mounted improperly? It was a pain to get on but at the end all the pins were pushed through properly like the manual said. At one point while putting the heatsink on I got it partially secured and and then had to remove it and immediately put it back on. Could that have been enough to screw up the thermal grease badly enough to cause this sort of problem? I've been running it with F@H for months continuously now, with the temperature presumably at 95°C the whole time. Is it likely I've caused some sort of permanent damage?
Earlier today, I downloaded Real Temp 2.6 on a whim and it told me my CPU's core temperature for both cores was hovering around 95°C (!!). Distance to TJMax varied from 1 to 0 (As the documentation explains, TJMax is the maximum safe operating temperature; at which the processor automatically throttles down clock speed) I turned off Folding@Home and it immediately declined, finally stabilizing at around 45-50°C. When I turned F@H back on it climbed up to the same as it was originally in a couple of minutes and stayed there until I quit the program. I downloaded Core Temp, which reports the same except about 10 higher. Temperature is actually reported by the sensor as distance from TJMax, so the discrepancy is just because Real Temp thinks it's 95°C and Core Temp that it's 105°C. For both, however, the processor shot right up to the limit when put under load with F@H. Obviously, something is very wrong.
The question is, what's wrong and what should I do? I checked and the fans are still spinning fine. Is the sensor wrong? Is the heatsink mounted improperly? It was a pain to get on but at the end all the pins were pushed through properly like the manual said. At one point while putting the heatsink on I got it partially secured and and then had to remove it and immediately put it back on. Could that have been enough to screw up the thermal grease badly enough to cause this sort of problem? I've been running it with F@H for months continuously now, with the temperature presumably at 95°C the whole time. Is it likely I've caused some sort of permanent damage?
Salvius on
0
Posts
Yet.
I do have cats. Checking the heatsink doesn't show any fur, but there's a bit of dust clumping at the center. Doesn't really seem like it would be enough to make the processor so much hotter than normal. I could try cleaning it out a bit to see if that helps. Any recommended way to do that?
It's just the top end for Macbooks. Nothing odd about it. I try not to let it stay there very long. Annoyingly enough it's just the iScrobbler app that causes it to do that for about 8 minutes each day.
OP, open the case (after it's been off, so it's cool enough to touch the heat sink), and feel if the heat sink has come loose or detached somehow.
It seems sturdy. What little I could wiggle it was just the motherboard bending slightly, it doesn't seem to move independently of the board at all. However, you gave me an idea. I ran up the temperature with F@H and touched the edge of the heat sink. It was just a bit warm. The passive Northbridge heatsink next to it was substantially warmer. I'm guessing that means that either heat takes a long time to propagate through it or it's not properly absorbing the heat from the chip at all.
I'm going to bed now, so I won't be answering questions for a while.
wunderbar's comment is also important...say you're fucking around in your case without grounding yourself and you zap something with static electricity. It may boot up and behave totally fine. That doesn't mean you haven't damaged something that is now going to deteriorate. Months later you could have much slower performance or a hardware failure and have no idea why. Needless to say, 95 is NOT an acceptable temperature.
PSN: TheScrublet
Was playing AoC last week and all of a sudden my FPS drops to 8. I figure it's because of a patch they had just pushed out, screwed with vid drivers and such right up until the PC just shuts off. Started it back up and BIOS says processor is at 89c.
When I opened the machine up, the 3 of the 4 crappy posts on the Intel stock heatsink and come loose from the motherboard. The fan and heatsink combo was literally just hanging there kind of touching the processor heatspreader.
Picked up a new heatsink and some thermal compound, making sure the heatsink was the type that BOLTS through the MB to a plate (i.e. doesn't rely on surface tension, nylon posts and hope to keep your processor cool).
Getting 36c idle and 44c when pushed.
PSN: Broichan
but they're listening to every word I say
I bought the Arctic Freezer 7 Pro from that article about a year ago, have had no issues whatsoever - I think I can agree that either of them will definately serve you well. Basically turned the idle temp from the crappy stock Intel cooler into the new 100% maxed load temp.
I doubt you've caused permanent damage if it hasn't affected performance, but that kind of thing could go at any moment if you leave it run at high temp. You will get metal fatigue from temperature changes, which eats away at the expected lifetime of your processor (so does leaving it run at high temp), so having the temperature spike from 40 C to 90 C often is bad. Generally, processors keep functioning long after you've replaced it because its too damn slow, I wouldn't worry to much if your correcting the problem now.
I know the smaller the architecture, the more risk there is at high temp, 45nm processors naturally stay cooler, but if they do heat up a lot they are far more likely to fail at the same temperature then say a 90nm or even 60nm.
Since we're on the subject, I'll be building an Intel when that new Nehalem memory architecture comes out late this year. I am not up on heatsinks AT ALL. What's good? I'm not looking to overclock, just trying to keep the system well cooled. I am NOT interested in having the more extreme heavy heatsinks on the board (in other words, definitely staying well within the Intel-specified weight ranges).
PSN: TheScrublet
Ah... phew! Thank goodness that's all it was though.
And those pushpins were why I switched back to AMD, despite them having inferior processors these days. AMD still uses the extremely reliable Heatsink Clip, which I've never had dislodge or come loose, and it always has ensured a solid connection between the chip and heatsink.
The last Intel I bought had their new pushpin system (those plastic posts). I thought I'd gotten it in... I'm pretty sure I did. Apparently not solid enough for moving though. I moved the computer in my car from my old apartment to my new one. Plugged it in, and the damn thing overheated nearly instantaneously. I opened the case, and the pins had come loose, allowing the heatsink to free-fall OFF of the processor entirely, hitting my video card and destroying it.
So thanks to those push pins, I lost a (pretty good) video card. When I got it re-attached, it still wouldn't boot... I'm still not entirely sure, but I think the processor itself fried. I'd about had it. I went that day to Fry's and picked up an AMD processor and board. I bit the bullet on performance for peace of mind. Haven't looked back since.
I bought a cheapo heatsink & fan for intel cpu's of this socket and it snapped so easily and firmly onto the board & cpu that it made the stock one such a joke.