Can bad RAM cause a CPU to overheat?

spainchaud

Reputable
Nov 25, 2015
4
0
4,510
I have a new computer from CyberPowerPC that has been crashing (WHEA_UNCORRECTABLE_ERROR, 124) when I play a MMO. It has an i7-6700K overclocked to 4.8GHz. I installed HWMonitor Pro and noted that core #1 was reaching temperatures in the range 80-90 deg C when loading a character into the game. I also noted that the BSODs seemed to occur at splash screens as opposed to heavy graphics activity. I removed one of my 8 GB sticks and saw that after repeated logins the temperature of core #1 never went over 57 deg C. I then swapped the RAM sticks and now I see a temperature of 87 deg during login. BTW, this memory passed an 18 hour long MemTest86. So I guess this stick has a problem, but I don't understand why RAM causes the CPU overheats and MemTest86 saw no problems. BTW, core #1 seems to run a good 10 deg hotter than the other three. Perhaps that is another bad sign?
 
Solution
1.5V is an dangerously high value for Vcore. Are you sure, can you post a screen shot of your BIOS?

I run both an overclocked i5 and i7 at 4.6 Ghz and under 1.3V because to get to 4.8Ghz requires about 1.38V and it's not so good to run at that voltage 24/7.

Is it at least an Adaptive Overclock and not an Override?
What Vcore are you running at? What cooler are you using? It is not unusual for cores to run ad different temperatures due to slight manufacturing differences. 10C is a lot though and the first thing that would occur to me is that the thermal paste needs to be applied more evenly.

In many games, less memory will reduce performance, so taking a stick out may slow the game down a bit, which reduces load on the CPU.

Going from 8Gb (2x4) to 4Gb (1x4) slowed things down by more than 10% when I did some benchmark tests.
 

spainchaud

Reputable
Nov 25, 2015
4
0
4,510
The temperature drop is not related to having only one stick of RAM instead of two. After I posted here I ran a test with two new sticks of RAM (total 16 GB). The peak temperature in core #1 reached 55 deg, same as for just one RAM stick.

CPU Vcore is 1.5V. On my invoice the water cooler is described as CYBERPOWERPC EX240 DC 420 RAYSTORM WATER COOLER KIT.

I am going to guess that the bad RAM was somehow causing the CPU to work harder. Possibly it has to make multiple requests to get the data it needs?

One other strange thing I am seeing. Last night, when running my temperature tests, I played the MMO in a window using about 2/3 of the screen, so I could also watch HWMonitor. This morning I am playing full screen and tabbing to HWMonitor when I swap characters. I now see a peak core #1 temperature of 75 deg. So about 20 deg higher. I have to go to work now, maybe I will do more tests tonight to help understand how CPU temperature is affected by different testing scenarios.
 
1.5V is an dangerously high value for Vcore. Are you sure, can you post a screen shot of your BIOS?

I run both an overclocked i5 and i7 at 4.6 Ghz and under 1.3V because to get to 4.8Ghz requires about 1.38V and it's not so good to run at that voltage 24/7.

Is it at least an Adaptive Overclock and not an Override?
 
Solution