How badly damaged is my cpu? (Damage in terms of long-term damage)

comanzo

Honorable
Jun 18, 2016
150
0
10,690
Hello Community.

This is going to be a long thread, so please bear with me. So, I have an i7-4790s, and wanted to test if the stock cooler would be good enough for cooling, since it's at stock settings. Now, you may say that doesn't need testing, as any stock cooler would be good for a cpu non-overclocked. However, my OEM case has bad airflow, as I only have 1 case fan running, minus the cpu fan, and the gpu fan. Therefore, one could make the argument that the stock cooler is no good for that particular situation.

I decided to test using IntelBurnTest. Now, keep in mind please, that at the time of testing, I wasn't as knowledgeable in pc hardware, and didn't know what the specs meant. Out of sheer ignorance, lack of knowledge, impatience and stupidity, I followed the TJ max specs, and told myself as long as the cpu didn't reach that, I was good to go. As you all know, IntelBurnTest is a cpu test that occasionally heats up for several seconds, and then cools off for another couple of seconds. The test probably ran for 1 minute, to 80 seconds tops. Out of that time, probably 50% of the time was when it was heating up.

The max temperature on the cpu core was 86 degrees Celsius. I know because I made a thread on this site talking about it. For those who want to look at it, I will post the link shortly. Now, to validate the results, I probably ran the test 2-3 times in a row. Why? As you know, the longer it's stressed, the hotter the temperatures get; therefore, doing it several times simulates temperatures after long periods of time. With that said, after doing it 2-3 times, and 50% of the time throughout each test, it reached the mid-80's, we can say the cpu was in the mid-80's for a good minute and a half, to two minutes tops.

Which brings me to this point. If the max temperature of ONE of the cpu cores was running at 86 degrees Celsius for a good 2 minutes straight, how badly damaged is the cpu? I ran the test a while back, meaning that this question is meant in terms of long-term damage, as short-term damage would've happened already. Supposedly the advice given for my cpu is to never reach above 80(79 if you want to be exact). Meaning that I went above the advice given by 7 degrees Celsius, for a good 2 minutes approximately. I did probably run this test at other moments too, but certainly not a lot of times such as this one moment. Sorry for the long post, and hopefully next time, I pay attention to what the cpu specs mean before doing such strenuous tests. Thanks. :)
 
Solution
So, you'll find some... general disagreement. In servers, it's normal to let the CPUs run at 85-95c and these chips are often serve critrical functions, and truck away for years before being retired. For mobile Haswell CPUs, that 100c figure is Tjmax, or max temp. Manufacturers allow laptop and mobile CPUs to run close to that with no apparent long-term issues. However desktop CPUs only have a "Tcase" temp published, which is the max temp the top of the heatspreader can get as spec'd by Intel. We can't measure this though, since there's no sensor there, and many on this forum take a very conservative view and suggest staying well back from throttling temp. They're the same physical CPU though. Also, on the new Kaby Lake chips, the only...
Zero damage.

That CPU can run at full speed, sustained, without damage, upwards of 90c.
Now this isn't advised - but it is fine for the chip.

The chip will throttle itself (slow down) to prevent over heating if it gets to that point.
If that doesn't solve it due to some other problem and temps continue to rise, the PC will eventually halt and shut down due to thermals.

You could run your CPU full load at 86degC 24/7 and it would probably still out live you.
 


Ok. That makes me feel better. I noticed some stuttering in some games, and was thinking it was due to this. No, it wasn't my gpu or wifi, so I automatically put the blame on this situation. Now I know it couldn't have been this.Thanks for answering so quickly. One last question, this never happened, but is just out of curiosity. Suppose my cpu did shut down due to thermals. Since that is one of the protocols for cpu protection, would there be damage in that situation? In other words, since it shut down due to overheating, even in that situation, there should be no damage correct, since it shut off in time?
 


Thanks for the quick reply. I always thought this was the reason why I had stuttering in my games. One last question. Since cpu's downclock or even shut off for thermal protection, if mine were to shutoff, there would be no damage correct, since it shut off in time? In other words, since it's one of Intel's ways of protection, it shouldn't cause damage since they put it low enough to shut off before damage occurs correct?
 


Ok. Got it, 100 for long-term stability, as well as extended periods of time, and 130 for short-term danger. One last question, are these temps. for my cpu(haswell), or do these temps. go for any architecture. Thanks for the quick replies.
 
So, you'll find some... general disagreement. In servers, it's normal to let the CPUs run at 85-95c and these chips are often serve critrical functions, and truck away for years before being retired. For mobile Haswell CPUs, that 100c figure is Tjmax, or max temp. Manufacturers allow laptop and mobile CPUs to run close to that with no apparent long-term issues. However desktop CPUs only have a "Tcase" temp published, which is the max temp the top of the heatspreader can get as spec'd by Intel. We can't measure this though, since there's no sensor there, and many on this forum take a very conservative view and suggest staying well back from throttling temp. They're the same physical CPU though. Also, on the new Kaby Lake chips, the only temp spec Intel has published ins Tjunction (max core temp), and it's also 100c.

So, make of it what you will. Running cooler will not hurt anything. There's strong evidence that at least short term use near 100c is harmless, and more evidence that in related CPUs, running close to throttle temp long-term is also not an issue, but you're not going to find clearcut instructions from Intel on the matter.
 
Solution
Alright. Thanks for the information. It kind of sucks intel is so ambiguous with desktop cpu's, as that's the one I am using. On the bright side, since the testing was well below 100 degrees celsius, and wasn't for a long period of time as it was less than 5 mins., my CPU is in the clear. Will note from now on to keep below throttling temps(100 degrees C) for long-term use/stability. Thanks again.