Is cpu degradation something that can be prevented from getting worse?

Toggle sidebar Toggle sidebar
V

viking78

Reputable
Jun 16, 2014
13
0
4,510
Sorry for the long post.

Ok, so my computer recently became infected with something that nothing seemed to remove, not even ComboFix so I decided to try for the first time a rootkit removal tool called GMER. To my horror, after a couple of minutes running, this GMER tool made all the 4 cores of my cpu (an i7-2600k overclocked to 4.4ghz) to jump from 65-70c to 98c (core temp literally reported "98c(?)", which I suppose means that it could be over 98c). After a few seconds I just switched off the pc the hard way as the pc seemed to have become clogged. At this point I thought it had to be some sort of bug, thinking that in a more realistic scenario, such an increase in core temperatures should have been much more gradual, so I decided to run GMER again with the exact same result as the first time, a 30-degree temperature jump on all cores (with core temp again reporting " 98c(?) " so I turned off the computer again as fast as I could. I don't think the cpu throttled on neither occasion eventhough I'm sure intel adaptive thermal monitor was activated in the bios.

Eversince, the cpu has been running without visible problems but I was concerned about degradation so I decided to run IBT on maximum (before the incident, I had my i7-2600k running at 4.4ghz with a manually-set voltage of 1.335v, the lowest voltage I could manage in order to pass at least 10 rounds of IBT on maximum, although it might have been 25 rounds what it could pass at 1.335v) and what I saw was that the cpu can no longer do 25 or 10 rounds of IBT on maximum, requiring 1.340v to pass the tests.

So it looks like the cpu did reach critical temps and that a slight degradation has occurred. What I wanted to know is if degradation is a process that can be stopped from becoming worse all the time or I'm basically screwed.

Thanks
 
Solution
P
Sort by date Sort by votes
V

viking78

Reputable
Jun 16, 2014
13
0
4,510
My idle temps are in the 40s, so they're fine. When I mentioned the jumps from 65-70c to "98(?)", I was referring to something that happens when the cpu is already on load...my temperatures are ok considering I've got an aftermarket air cooler (80-82c max when running Prime95 small FFTs or IBT on maximum).

But I did get unlucky with the silicon lottery, not the best overclocker my i7-2600k. I should be getting at least 4.5ghz stable at 1.335v, not 4.4. Also, this was my first real overclock and since I found the offset method a bit complicated, I went for the manual set voltage way. The thing is, software like CPUID HWMonitor or HWiNFO64 report a max voltage of 1.360v, despite being set in the bios at 1.335v. It also seems to go under 1.335v too (maybe down to 1.320v) although it does seem to average around 1.342v as reported by HWiNFO64. I have also tried turning off internal pll overvoltage and that seems to make the voltage fluctuate a little bit less, although it still fluctuates.

Is it supposed to be like that?
 
Upvote 0 Downvote
P

Pinhedd

Champion
Aug 7, 2007
10,342
20
56,965


Intel Burn Test is designed to put the CPU under synthetic loads that maximize thermal stress. I am not aware of any other application that is capable of causing that kind of power draw. If IBT doesn't raise the CPU to a certain temperature, nothing else will.

Given that root kits work by interfering with low level system operations, it's plausible that any tool designed to remove root kits may cause problems with low level operations. Core Temp is also known to have issues, so it's most likely just a false reading.
 
Upvote 0 Downvote
Solution
V

viking78

Reputable
Jun 16, 2014
13
0
4,510
I also wasn't aware of any application that could push the cpu nowhere near IBT on maximum...that's until I played Crysis 3 with my i7-2600k. Early in the game there's a level completely covered in foliage plants which are constantly moving in the wind and, while the physics on maximum are simply amazing, they literally murder my cpu...no other cpu-intensive game (ie watchdogs or battlefield 4) heats up my cpu like that.

Just so you get an idea, when I played the game on that level, the max temp on the hottest core was only 2 degrees cooler than the max temp on that same core (always the hottest one) when running Prime95 with small FFTs. Also, the max temp on the hottest core when running IBT on maximum is only 2 degrees hotter (and I've read numerous times in forums that IBT gets the cpu 10-15 degrees hotter than Prime95, lol) than than the max temp on that same core when running Prime95 with small FFTs.

So yeah, that game sometimes stresses the cpu virtually like a torture test program.


What concerns me a bit as well is that the max vcore reported by software such as CPUID HWMonitor or HWiNFO64 is always 0.020-0.028v higher than what I've actually set it to in the bios (at the moment I have it set to 1.300v but CPUID HWMonitor is reporting a max vcore of 1.328v). I thought it might have been having pll internal overvoltage activated in the bios but deactivating makes no difference. This concerns me somewhat when the cpu is at 4.4ghz since I cannot manage less than 1.335v but the max vcore reported will be 1.360-1.368 with an average of 1.341v as reported by HWiNFO64. If I want 4.5ghz, the lowest I can manage is 1.355v but that means the cpu will spend a lot of time at 1.360-1.365v, with peaks of 1.370-1.378 and Intel apparently recommends a maximum voltage of 1.355v for 24/7 use.
 
Upvote 0 Downvote
P

Pinhedd

Champion
Aug 7, 2007
10,342
20
56,965


Yeah Crysis 3 is nuts. Keep in mind that your GPU will be going full blast as well, so it will affect the temperature of the CPU indirectly.

As to the Vcore thing, don't worry about it. That's normal behaviour for many motherboards, especially Asus motherboards.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom