Processor Gets Stuck at Lowest Multiplier

Darren

Darren

Moderator
Staff member
Hey all. This isn't really a problem for me but more of just something I stumbled across today. I've got an 8320 in the machine in my sig. I've had it for a while and had no problems. One day I wanted to overclock the chip and see how far I could get it. This was a while back and I think I was getting stable at around 4.2GHz, but could push it higher to around 4.4GHz or so. I'd get BSOD from prime, but games were okay. Temperatures weren't even breaking in to the 60's. When I would do some of my stress testing I would notice my processor would read in HWMonitor as being 206 degrees Celsius, which is downright alarming. I would be smelling smoke at that point. After that day I just knocked everything back to default and have been that way ever since.

Today I was curious what my load temps were and fired up Prime at stock clocks and ran it. After a couple minutes my temp shot up again. CPU-Z also noted something. My processor was running fine at first but when the temperature sensor spiked the processor downclocked to 1.4GHz with a multiplier of 7 at 200MHz. I ran Skyrim and sure enough, game was unplayable. CPU is currently locked at 1.4GHz. Restart of the machine will fix this issue I'm sure as I get the feeling this was probably going on when I was overlclocking and just didn't notice this.

During gaming and regular use I NEVER have any issues at all. This just so happened after running Prime. What's going on? I'm more curious than anything. I'm guessing my motherboard is just kind of crappy, which wouldn't surprise me, it was cheap and needed a BIOS update to even support Piledriver as it's over 3 years old now.

TL;DR Prime95 makes my temp sensor incorrectly read over 200 degrees Celsius and my Processor is stuck at a downclock of 1.4GHz.

Edit: Also, HWMonitor is still reading me at 206 degrees Celsius right now at idle.
Edit 2: Restart has everything back to normal.
 
Last edited:
spynoodle

spynoodle

Active Member
You were probably being downclocked because your CPU was thermal throttling. I don't know why your sensor would read over 200 degrees C, because there is basically no way that your CPU could handle that, but either way it seems like your sensor was attempting to report a very high temperature. Perhaps it was just misreading because the temperature was so high. Either way, you probably need to reduce your overclock. That burning smell can't be good, either; there could be a problem with the capacitors on your motherboard.
 
Darren

Darren

Moderator
Staff member
spynoodle said:
You were probably being downclocked because your CPU was thermal throttling. I don't know why your sensor would read over 200 degrees C, because there is basically no way that your CPU could handle that, but either way it seems like your sensor was attempting to report a very high temperature. Perhaps it was just misreading because the temperature was so high. Either way, you probably need to reduce your overclock. That burning smell can't be good, either; there could be a problem with the capacitors on your motherboard.
Click to expand...

That's what's weird though. This was at stock clocks. The CPU temp sensor was reading 45 degrees Celsius then would immediately jump. I've got Arctic Silver 5 with a 212, it's not overheating. Also I never said I had a burning smell, but that I would have smelled something if it did get hot.
 
Last edited:
Okedokey

Okedokey

Well-Known Member
Darren said:
That's what's weird though. This was at stock clocks. The CPU temp sensor was reading 45 degrees Celsius then would immediately jump. I've got Arctic Silver 5 with a 212, it's not overheating.
Click to expand...

You've probably damaged the CPU mate. RUnning it at 200oC + for even a moment kills it. How are you testing temp when you run prime?
 
Darren

Darren

Moderator
Staff member
Okedokey said:
You've probably damaged the CPU mate. RUnning it at 200oC + for even a moment kills it. How are you testing temp when you run prime?
Click to expand...

There is no way it really hit that temperature, I would have smelled it burning. Also it wouldn't immediately jump that much that fast. It was instant.

HWMonitor.
 
Okedokey

Okedokey

Well-Known Member
A lot of certainty you seem to have. A CPU without proper connection to the heatsink can increase in temperature very very quickly, essentially instantly.

Download PC WIzard from my sig and tell me what that says.
 
Darren

Darren

Moderator
Staff member
Okedokey said:
A lot of certainty you seem to have. A CPU without proper connection to the heatsink can increase in temperature very very quickly, essentially instantly.

Download PC WIzard from my sig and tell me what that says.
Click to expand...

Do you seriously think my motherboard would be in any functioning state if the CPU stayed at the temperature for more than a few seconds? I'd leave Prime running for a while when it was reading at that temperature and nothing happened. The sensor is wrong, I'm certain of that. I'm just more curious as to what's going on. Board instability is the only thing that really makes sense.

Furthermore I know how to install a HSF and can assure you it's installed perfectly fine. The temperature is not my concern, I'm wondering how or why the chip is getting locked to a lower speed. Also the first time I noticed the temp get that high when I was doing OC testing a few months back I immediately turned it off. I've since had that high temp throw several times and it's clearly not that warm. Sticking my hand above the exhaust fans has air flow no warmer than normal.

What info are you wanting from PC Wizard..?
 
Darren

Darren

Moderator
Staff member
Alright as you can see here both HWMonitor and PCWizard are reading incorrectly. Prime is still trucking along and no showing any problems and processor usage is maxed according to Windows. Clock speed is at 1.4GHz according to CPU-Z. Can confirm my computer is not melting so the 206 is clearly not right.

H96UD19.png
 
FuryRosewood

FuryRosewood

Active Member
i cant imagine a cpu pulling that much temp...without melting the damn motherboard. lol thats oven level temps...
 
spynoodle

spynoodle

Active Member
Given everything you've said, my guess would just be that somehow the temperature sensor got damaged, and now the CPU thinks that it has to throttle when it really doesn't. I'm not sure if it's even possible to fix a broken temperature sensor, so your only option would be to disable thermal throttling somewhere in the BIOS. I'm not sure if this is even possible with your board/proc, but I know for sure that it is with my old board.

Sorry I misread your first post, by the way; I thought you were saying that you actually smelled burning. :D
 
Darren

Darren

Moderator
Staff member
Nah. I'm not going to really change anything either since it only does this in prime.
 
Okedokey

Okedokey

Well-Known Member
Actually it is because your motherboard really wasn't designed for it. I would physically touch the mosfets on the power side of your motherboard, I bet they're cooking when you ramp it up.

Blow air at this area

970A-G4681205.jpg
 
Darren

Darren

Moderator
Staff member
Reading through the exchange that the guy had with MSI support I'd be more inclined to say AMD is at fault for this more so than MSI. Either way it really doesn't matter for me because I never hit 100 percent load anyway. Thanks for the input guys.

When I do upgrade I'll definitely be getting a stronger motherboard. Probably a nice Intel chip to boot. :D
 
StrangleHold

StrangleHold

Moderator
Staff member
See if you can disable that APS in the bios. Might work good with a lowend CPU. But a 125W I just don't see how ramping it up and down could help. Plus like Okedokey said, those mosfets are the older design. They have 4 on each phase but I bet they are still getting pretty hot.
 
You must log in or register to reply here.
Top