How to stress test my " possibly damaged cpu"?

Ali_76

Distinguished
Dec 11, 2015
59
0
18,640
So intel gave me a replacement 6600k after my old one died. I looked at the back when i was applying my thermal paste and ive seen that couple of the golden plates had an obvious orange dots in the centre almost as if it was thermally damaged.
photo http://imageshack.com/a/img924/3654/XJNIVL.jpg

if you look at the top row of the cpu golden plates, you can see an orangish color. The cpu works great and its not overclocked yet adn i was wondering if there are safe stress testers out there to know if my cpu is working or not or atleast not going to fail in the next few months. I ran prime 95 and my temps went from a stable 35 degrees celsius to almost 70-71 and thats when i stopped the test.
 
Solution


The H60 doesn't actually use pure water, it is filled with a glycol mixture much like antifreeze. But the premise is the same. Yes, use the CPU Opt header for the pump, and be sure it is running at full speed. I believe that is around 4500 RPM. Then plug the 4-wire PWM rad fan into the 4-wire PWM CPU header.

clutchc

Titan
Ambassador
P95 is as good a stress test as there is. Maybe add Intel Burn Test to your collection, too.
Is your cooler working as it should? Fan not bogged down?
These are the recommended safe temps for the Q6600 depending on which you have:
5°C - 62.2°C (Stepping B3)
5°C - 71°C (Stepping G0)
 

Ali_76

Distinguished
Dec 11, 2015
59
0
18,640


 

Ali_76

Distinguished
Dec 11, 2015
59
0
18,640
My cooler is h60 water cooler and my thermal paste is arctic silver 5, the cooler works fine i guess, just making fan sounds here and there. and btw idk what q6600 , i have a core i5 6600k skylake cpu
 

Ali_76

Distinguished
Dec 11, 2015
59
0
18,640
no the old 6600k before it went defective was running at 30 idle ( 5 less than now) and never got it to stress or full load as i never had a graphics card. The h60 is attached to the fan blowing air from outside into the pump so >^ . i lost one of the screws for the fan so i have 4 attached so i dont see a problem there. My case is a carbide spec m2 a matx case with a " non modular psu" so everything is everywhere. could that be the issue. and formy thermal paste as5 , i applied it using the line method and i maybe did a slightly too much that recommended but its working fine rightn now, 28 degrees celsius at performance fan speed
 

Ali_76

Distinguished
Dec 11, 2015
59
0
18,640


 

Ali_76

Distinguished
Dec 11, 2015
59
0
18,640
oh ok , so the fan for the liquid cooler is attacked to the SYS FAN1 on my z170 MX gaming 5. and the wire on the waterblock is connected to CPU_FAN but there is a CPU_OPT that the manuel says is only for water coolers, idk if h60 is considered a water cooler
 
i really wouldn't worry on thermal damage - intel's cpus have a 100C thermal throttle limit built in, so it will decrease the load for a few seconds to bring the temp back down. On my first build, i was using the stock intel cooler and relying on Asus's AI Suite performance utility which was showing cpu temps never exceeding 67C.

i rendered videos for 3+ months, but noticed the computer would slow down at times. Then i saw a thread where someone complained the Asus performance utility was corrupt and showing heat or temp values way off. So i downloaded RealTempGT monitoring utility and it was showing 100C - cpu would run 96C to 100C then fall right back to 96C, and climb right back to 100C, and repeat endlessly, while rendering videos. Average video file took back then 75-85 minutes to render, and Handbrake is a core hog, showing 98-99% load in all four cores. I tested with a couple of other utilities including Intel's Extreme Tuning Utility, and all showed the same high temps kissing the throttle limit. So for 3 months i had been running my cpu flat out at the temp limit.

I deleted Asus's utility, went thru the system to see what in the hell was wrong, and found the intel cooler had one pin that would not stay locked, meaning the heatsink was not in full contact with the cpu. But needless to say i was worried i had damaged my cpu.

I upgraded the cooler, am now running temps that max at 63C, and when i run a benchmark using intel's ETU and then let the utility upload the results to HWBOT to compare to other users, the benchmark consistently places in the top 5% of other users running the same mobo and cpu - and this is two years after discovering the Asus utility was lying to me.

Those same videos now render in 35-45 minutes, and i generally do 5-10 a week.

Plus, i doubt the discoloration was from cpu temps - i work with metal, and it takes, depending on the alloy, 400-900 degress F to discolor stainless steel. Additionally, i've never seen stainless discolor orange from heat, usually blue and some black. Those orange spots are probably from the heat paste, maybe there was some contamination in it or one of the compounds in it turn orange naturally from heat.

FWIW
 

Codrru

Reputable
Aug 20, 2015
103
0
4,710
sorry to chime in late, i agree with the original ralph i wouldnt be worried about it. btw just to address the fan plugs, the opt. one is in case you have 2 cpu fans like on my noctua cooler, or a liquid cooler. (which your h60 is) so feel free to put it there. honestly though if your really concerned about it. the prime 95 test is there for a reason, use it for an hour no o.c. and if it fails send it back to intel as defective but the previous posters are correct your tj max is pretty high up there so i wouldnt be concerned about 70's c remember this test is meant to push it to its max.
 

clutchc

Titan
Ambassador


The H60 doesn't actually use pure water, it is filled with a glycol mixture much like antifreeze. But the premise is the same. Yes, use the CPU Opt header for the pump, and be sure it is running at full speed. I believe that is around 4500 RPM. Then plug the 4-wire PWM rad fan into the 4-wire PWM CPU header.
 
Solution