[SOLVED] CPU fried by running a script

Jan 1, 2022
3
1
15
My CPU died while running a heavy load script.
I built this PC about 10 years ago, MSI 760gm e-51 (fx) MB, 8350 AMD CPU. Only problems I ever had was overheating (shutdowns) while playing games, which I eventually fixed by adding a liquid cooler to the CPU. It ran fine for the past couple years, my only concerns were optimizing graphics to get a better fps in Skyrim SE. One avenue of that process is LODGEN, which optimizes distant viewing among other things. The last process I tried was generating occlusion files, where the script goes through the game files and generates new/different occlusion files to make the game 'better'. I don't claim to understand the ins and outs and the whys and hows. So let me get to the point.
While running the script, my pc shutdown, which hadn't happened since I installed the liquid cooling. So, I rebooted and turned on my hardware monitor and re-ran the script. This is where you'll realize I am an idiot, and this will be proven again in the near future :-0. So, I ran the script again, carefully monitoring the CPU core temp, as the process went from 50--60% done the CPU temp was rising above 60c, then to 70c, then to 80c and it was 95% done, so I foolishly thought I could eek out the process, but the pc shutdown just before it finished. I waited a minute and tried to reboot, but all I got was a soft 'pop' sound and couldn't restart into POST. So I figured I fried it and I ordered a replacement 8350 CPU. I installed it and the pc booted up just fine and everything was back to normal. And then I got really stupid, and decided I'd try running that script again, but this time I'll watch the CPU temp and cancel the process if the temp got over 70c. But, this time it got about 50% done, the CPU temp never got over 45c and the pc shutdown again. So, I pressed the start button and I hear that soft 'pop' sound again and the pc won't even POST again. Same exact scenario, but it apparently had nothing to do with heat.
What could possibly cause this? Is there a way to reset or fix the CPU? Was there some voltage setting captured in CMOS that fried it on reboot? How could running a script do this? Any insight or help is appreciated.
 
Solution
1)Motherboard not built to handle those 125w+ rated chips.
With the kind of power the FX-8350 is capable of drawing, the MSI 760gm e-51 was comparatively naked. It was only a matter of time before it started to crap itself.

2)AIOs do not provide as much direct airflow over the mobo VRMs as air coolers do. In turn, the VRM sees higher operating thermals. So you helped the cpu, but kinda left the already bare mobo hanging.

3)Using the wrong app to read FX cpu core temperatures. You can end up in a scenario where the cpu is running too hot or thermal throttling, and you'd not notice it, because you're looking at it from an ambient UP TO 100C range, which isn't accurate.
The onboard sensors on those cpus reported 'distance to thermal...

Phaaze88

Titan
Ambassador
1)Motherboard not built to handle those 125w+ rated chips.
With the kind of power the FX-8350 is capable of drawing, the MSI 760gm e-51 was comparatively naked. It was only a matter of time before it started to crap itself.

2)AIOs do not provide as much direct airflow over the mobo VRMs as air coolers do. In turn, the VRM sees higher operating thermals. So you helped the cpu, but kinda left the already bare mobo hanging.

3)Using the wrong app to read FX cpu core temperatures. You can end up in a scenario where the cpu is running too hot or thermal throttling, and you'd not notice it, because you're looking at it from an ambient UP TO 100C range, which isn't accurate.
The onboard sensors on those cpus reported 'distance to thermal margin, or TJ Max'. In the case of the FX-8350, which has a margin of 61C, that range is 61C DOWN TO 0, or even negative values.
AMD Overdrive and hwinfo can read them correctly.



2 out of 3, my guess would be that you were pushing the mobo too hard.
The hardware is designed to throttle down if things get too hot or too much power is being drawn, but the mobo here stands out as being under-specced for the cpu being used. Plus, it was pretty long in the tooth; 10 years is a nice run.
 
Solution

Rogue Leader

It's a trap!
Moderator
Agree with Phaaze88 the first time around it popped a VRM, the second time around it popped one or more additionally. That board is a 4+1 so it doesn't even have the ability to fire up the CPU with minimum 2 burnt. Aside from the fact that board had no business running an 8350 to begin with.

Also while you were reading the temps incorrectly, in reality if sensors are outputting numbers like that you're well beyond the thermal limit. so the CPU fried anyway.
 
  • Like
Reactions: rainbowsandpuppies
Jan 1, 2022
3
1
15
Well then, I guess I was working on borrowed time all this time. 10 years is certainly a nice run, I had no idea the 8350 was too much for the MB. Derp. I guess I put too much trust into what is possible vs. what is practical. CPU-Upgrade: MSI 760GM-E51 (FX) processor support and specifications has it listed for instance. Lesson learned, trust less; research more. You mention the VRMs, am I to assume one of them dying was the soft 'pop' sound I heard? And now I am 2 down, meaning even with a new CPU the MB is dead? I was planning on getting another CPU, but what you say (as I understand it), I shouldn't and just get a new MB and CPU? Anyway, thanks for the feedback Phaaze88 and Rogue Leader, much appreciated!
 

Rogue Leader

It's a trap!
Moderator
Well then, I guess I was working on borrowed time all this time. 10 years is certainly a nice run, I had no idea the 8350 was too much for the MB. Derp. I guess I put too much trust into what is possible vs. what is practical. CPU-Upgrade: MSI 760GM-E51 (FX) processor support and specifications has it listed for instance. Lesson learned, trust less; research more. You mention the VRMs, am I to assume one of them dying was the soft 'pop' sound I heard? And now I am 2 down, meaning even with a new CPU the MB is dead? I was planning on getting another CPU, but what you say (as I understand it), I shouldn't and just get a new MB and CPU? Anyway, thanks for the feedback Phaaze88 and Rogue Leader, much appreciated!

Yes very much do not put another CPU in that board unless you like lighting money on fire.

It was fairly common knowledge around that generation that there was minimal control behind what CPUs were allowed to work with what boards, so basically almost any AM3+ board can boot almost any AM3+ CPU, but in terms of being able to fully power the CPU or overclock it the difference between a proper board and not was massive. This was perpetuated by shops selling prebuilts with crap boards and 8350 CPUs, and even places bundling that CPU with inadequate boards.

Needless to say even the cheapest AMD Ryzen or Intel i3 you can buy right now demolishes that 8350. No reason to throw good money after bad.