Hello all! I've always used these forums as an excellent resource, and recently ran into a concern I thought the community could help with.
I recently pieced together a budget rig over the course of a few months snagging some parts on sale. For reference, this is what I'm working with:
AMD FX-6300 (stock settings)
Cooler Master Hyper 212 EVO (w/ Arctic Silver Arctic Alumina)
Gigabyte GA-970A-UD3P (stock settings)
Fractal Define R4 - additional 140mm fan added to front panel for 3 total; 2 in (front), 1 out (rear)
I had intentions to mildly overclock, though I have had no experience with it in the past. At the time of this post and the event to be detailed I have not changed any BIOS settings in an effort to overclock.
I completed putting the box together with my son over the weekend and proceeded to install Windows 8.1 (I had a key gifted to me, thought I'd see what all the fuss was about). I also booted with the case open to make sure all fans were turning. After which, I rebooted and entered BIOS, and noticed my CPU temperatures would jump from 27C to 34C fairly erratically. I then read some posts about AMD CPU temperature sensors being wonky at low readings sometimes and decided I should stress test it to both burn in and confirm that my temperatures were acceptable.
So Monday night I loaded Speedfan and Prime95 from a USB and set to work. First, it was difficult with Speedfan to tell what was what (I understand I should probably be using HWMonitor or something else, and probably will when I decide to move forward again). I found a post from another user with the exact same motherboard and processor with both Speedfan and HWMon pulled up, and it seemed to verify that Temp 2 was the socket temp and Temp 3 was the CPU temp.
At any rate, everything seemed fine. Idle was in the low 20s. Loaded for a few hours, the CPU seemed to hold around 32C and the socket around 40C. Since I was about to go to bed, I figured I would leave prime running overnight, just to make sure there were no errors or problems.
The following morning I was still around 33-34C on the CPU, maybe 41-42C on the socket. Glad that I was getting a consistent reading now, I headed to work. I was only supposed to work a half day, so I thought I'd be getting about 16 hours of prime in, and I'd move on with overclocking or what have you that afternoon.
Well, work had a catastrophe, and I ended up there until later than expected, and arrived home at about 20 hours of stress testing (I've tested for longer than this in the past and wasn't terribly concerned). After returning home I find that my Temp 2 and Temp 3 are 91C and 79C respectively! Alarmed, I immediately stopped the test (which had no errors or warnings), at which point the temps began to drop to around 50C.
After some investigation, I found that my CPU fan (included with the 212 EVO) had failed at some point. I then shut down the PC, restarting only a couple of times to verify that (a) the CPU fan did not work in any other fan header (it did not) and (b) that other known working fans did work in the CPU fan header (they did).
Luckily, I don't believe the CPU ran for too terribly long at this temperature because the mean temp for Temp 3 was still listed as 34.9C, max at 79C. (Speedfan's chart only goes back 15 minutes so I could not verify) Doing a little bit of raio math (which is at best a rough estimate), it seems the CPU ran like this for about one hour.
TL;DR: My FX-6300 ran a stress test for an unknown (but probably around an hour) amount of time with a Hyper 212 EVO heatsink but no CPU fan running to a maximum of 79C (I think, see questions).
I guess my concerns are multiple, but here they are:
Am I correct in my assumptions of the Temp 2 and Temp 3 being socket and CPU respectively?
Should I be concerned about the health of my CPU (or other hardware) after this incident?
Does this warrant me removing the heatsink and cleaning/reapplying thermal compound?
I recently pieced together a budget rig over the course of a few months snagging some parts on sale. For reference, this is what I'm working with:
AMD FX-6300 (stock settings)
Cooler Master Hyper 212 EVO (w/ Arctic Silver Arctic Alumina)
Gigabyte GA-970A-UD3P (stock settings)
Fractal Define R4 - additional 140mm fan added to front panel for 3 total; 2 in (front), 1 out (rear)
I had intentions to mildly overclock, though I have had no experience with it in the past. At the time of this post and the event to be detailed I have not changed any BIOS settings in an effort to overclock.
I completed putting the box together with my son over the weekend and proceeded to install Windows 8.1 (I had a key gifted to me, thought I'd see what all the fuss was about). I also booted with the case open to make sure all fans were turning. After which, I rebooted and entered BIOS, and noticed my CPU temperatures would jump from 27C to 34C fairly erratically. I then read some posts about AMD CPU temperature sensors being wonky at low readings sometimes and decided I should stress test it to both burn in and confirm that my temperatures were acceptable.
So Monday night I loaded Speedfan and Prime95 from a USB and set to work. First, it was difficult with Speedfan to tell what was what (I understand I should probably be using HWMonitor or something else, and probably will when I decide to move forward again). I found a post from another user with the exact same motherboard and processor with both Speedfan and HWMon pulled up, and it seemed to verify that Temp 2 was the socket temp and Temp 3 was the CPU temp.
At any rate, everything seemed fine. Idle was in the low 20s. Loaded for a few hours, the CPU seemed to hold around 32C and the socket around 40C. Since I was about to go to bed, I figured I would leave prime running overnight, just to make sure there were no errors or problems.
The following morning I was still around 33-34C on the CPU, maybe 41-42C on the socket. Glad that I was getting a consistent reading now, I headed to work. I was only supposed to work a half day, so I thought I'd be getting about 16 hours of prime in, and I'd move on with overclocking or what have you that afternoon.
Well, work had a catastrophe, and I ended up there until later than expected, and arrived home at about 20 hours of stress testing (I've tested for longer than this in the past and wasn't terribly concerned). After returning home I find that my Temp 2 and Temp 3 are 91C and 79C respectively! Alarmed, I immediately stopped the test (which had no errors or warnings), at which point the temps began to drop to around 50C.
After some investigation, I found that my CPU fan (included with the 212 EVO) had failed at some point. I then shut down the PC, restarting only a couple of times to verify that (a) the CPU fan did not work in any other fan header (it did not) and (b) that other known working fans did work in the CPU fan header (they did).
Luckily, I don't believe the CPU ran for too terribly long at this temperature because the mean temp for Temp 3 was still listed as 34.9C, max at 79C. (Speedfan's chart only goes back 15 minutes so I could not verify) Doing a little bit of raio math (which is at best a rough estimate), it seems the CPU ran like this for about one hour.
TL;DR: My FX-6300 ran a stress test for an unknown (but probably around an hour) amount of time with a Hyper 212 EVO heatsink but no CPU fan running to a maximum of 79C (I think, see questions).
I guess my concerns are multiple, but here they are:
Am I correct in my assumptions of the Temp 2 and Temp 3 being socket and CPU respectively?
Should I be concerned about the health of my CPU (or other hardware) after this incident?
Does this warrant me removing the heatsink and cleaning/reapplying thermal compound?