Board dying or CPU?

B

bmxfelon420

n00b
Joined
Jan 22, 2006
Messages
52
Specs:
6700k
MSI Z170A Gaming
MSI GTX1080
Gskill DDR4 3200
PCP&C 1200w PSU

Hello, was just trying to play a game today and my PC power cycled itself in the middle of it. Checked, getting a bunch of WHEA-Logger event logs for various stuff. Mostly "Cache Consistancy" but I think that's just because it throws one for every CPU thread and they show more than the others. Saw a Memory, Processor, and PCI-Express in there too. I'm thinking the board is going bad, 2-3 times since I got it (I think 3 years now?) It goes nuts and wipes out all of my bios settings. The last time it happened I had to mess with it to not fail on reboot every single time. Now it started doing this. Turned it back down to stock/2133 RAM to see if the event log errors stopped, no luck.

Anyone have thoughts? Kinda hoping it isnt the CPU, these things are still stupidly expensive. If that's bad and they wont warranty it I might as well just get a new board/cpu.
 
My guess is that PSU has some age on it. If it's giving up the ghost, I would imagine it could send voltage flux that could result in scatter-data. That being said, you said the board has been flaky before - so that automatically makes it suspect.
 
Joust said:
My guess is that PSU has some age on it. If it's giving up the ghost, I would imagine it could send voltage flux that could result in scatter-data. That being said, you said the board has been flaky before - so that automatically makes it suspect.
Click to expand...
I have a brand new EVGA 1000w spare, I can test this if need be.
mda said:
Try a new CMOS battery? I've seen weirder things happen when the CMOS battery is dead.
Click to expand...
That's worth a try, the machine is generally on 24/7 so maybe it's causing an issue.

I'm going to go down to one of the stores and see if they have any z170 boards in stock just in case, I can always pull the CPU and test it out real quick. I have a spare set of RAM too so I can eliminate that as well if needed. I just dont have a board/CPU spare at the moment.
 
Are you OC'ing at all?
I got some WHEL errors myself even at stock when I wasn't pushing enough voltage to the CPU.
 
Spartacus09 said:
Are you OC'ing at all?
I got some WHEL errors myself even at stock when I wasn't pushing enough voltage to the CPU.
Click to expand...

It isn't uncommon for overclocked or even previously overclocked components to need a bit more voltage as they age. I've seen it several times in my own and friend's machines. You don't necessarily need to swap out the power supply to exclude it as an issue. Do some load testing with Prime95 and Furmark while monitoring voltages. If the voltages are more than +-10% nominal you may have an issue.
 
Spartacus09 said:
Are you OC'ing at all?
I got some WHEL errors myself even at stock when I wasn't pushing enough voltage to the CPU.
Click to expand...
I tested both overclocked and not, with varying voltages. I ran for 2 years stable at 4.6/1.35 vcore, noticed this power off issue recently and tested between 1.35 and 1.4v fixed as well as equivalent offset values. Offset was not stable for some reason though; I also set the CPU back to stock and disabled XMP to see if it affected the frequency of the errors, and thus far it has not.

kirbyrj said:
As a general rule of thumb, CPU's rarely die and it's usually the board's fault if the choice is between the two.
Click to expand...
I agree and have seen this 99% of the time, I did have an 1800x a month ago though that was bad. Swapped everything else first since it was the last thing I suspected.

Zareek said:
It isn't uncommon for overclocked or even previously overclocked components to need a bit more voltage as they age. I've seen it several times in my own and friend's machines. You don't necessarily need to swap out the power supply to exclude it as an issue. Do some load testing with Prime95 and Furmark while monitoring voltages. If the voltages are more than +-10% nominal you may have an issue.
Click to expand...
I'll fire up the monitoring, I wanted to see what the LLC did in Mode 1 anyways because I suspect it's going way too high, and I may need to turn down the Vcore if I change the LLC curve. I typically dont have voltage monitoring running all the time.
 
bmxfelon420 said:
Specs:
6700k
MSI Z170A Gaming
MSI GTX1080
Gskill DDR4 3200
PCP&C 1200w PSU

Hello, was just trying to play a game today and my PC power cycled itself in the middle of it. Checked, getting a bunch of WHEA-Logger event logs for various stuff. Mostly "Cache Consistancy" but I think that's just because it throws one for every CPU thread and they show more than the others. Saw a Memory, Processor, and PCI-Express in there too. I'm thinking the board is going bad, 2-3 times since I got it (I think 3 years now?) It goes nuts and wipes out all of my bios settings. The last time it happened I had to mess with it to not fail on reboot every single time. Now it started doing this. Turned it back down to stock/2133 RAM to see if the event log errors stopped, no luck.

Anyone have thoughts? Kinda hoping it isnt the CPU, these things are still stupidly expensive. If that's bad and they wont warranty it I might as well just get a new board/cpu.
Click to expand...

If anything has failed, its likely the motherboard or even memory. CPU's rarely go bad.

mda said:
Try a new CMOS battery? I've seen weirder things happen when the CMOS battery is dead.
Click to expand...

I have worked on thousands of machines and I've only seen a handful of them die. I've never seen anything beyond the clock not being right and the system not holding its settings in BIOS.

Spartacus09 said:
Are you OC'ing at all?
I got some WHEL errors myself even at stock when I wasn't pushing enough voltage to the CPU.
Click to expand...

Generally at stock settings on a good motherboard, you shouldn't have to increase voltages beyond the automatic setting. However, this isn't always the case. For example, we used to have to set the CPU input voltage on all X99 motherboards to .90v to .95v. Some systems defaulted to .80v or even lower and that was never stable at stock settings. Corsair RAM as another example, often needs to be set to 1.36v or higher to run at 1.35v.

kirbyrj said:
As a general rule of thumb, CPU's rarely die and it's usually the board's fault if the choice is between the two.
Click to expand...

Agreed. After having serviced well over a thousand machines as a service tech and many thousands more in the IT industry, I've seen probably fewer than a dozen genuinely bad CPU's from all vendors combined. That said, I did see one case where a CPU was modestly overclocked and degraded to a point where it had to be overvolted to run at stock speeds. Eventually, it had to be downclocked and overvolted to run at all before dying. Degredation like that I've seen only a handful of times. Its even rarer than actual outright CPU failures.

Zareek said:
It isn't uncommon for overclocked or even previously overclocked components to need a bit more voltage as they age. I've seen it several times in my own and friend's machines. You don't necessarily need to swap out the power supply to exclude it as an issue. Do some load testing with Prime95 and Furmark while monitoring voltages. If the voltages are more than +-10% nominal you may have an issue.
Click to expand...

Sometimes, this is the case. Although, in my experience when that happens failure isn't too far down the line. I absolutely agree on the power supply. You don't need super precision. The GIGABYTE utilities or its UEFI BIOS monitor would be good enough. If you really want to watch it under load, a cheap multi-meter can get the job done. You just need to ballpark it.

Joust said:
My guess is that PSU has some age on it. If it's giving up the ghost, I would imagine it could send voltage flux that could result in scatter-data. That being said, you said the board has been flaky before - so that automatically makes it suspect.
Click to expand...

When a PSU fails, it normally browns out or shuts down suddenly. A brown out leads to random reboots in most cases. I've rarely seen a PSU cause instability in a system. In fact, I've seen it a couple of times after replacing a few hundred PSU's in various machines over the years. And that was on my test bench after running a PSU on my test bench for about 10 years.
 
Dan_D said:
When a PSU fails, it normally browns out or shuts down suddenly. A brown out leads to random reboots in most cases. I've rarely seen a PSU cause instability in a system. In fact, I've seen it a couple of times after replacing a few hundred PSU's in various machines over the years. And that was on my test bench after running a PSU on my test bench for about 10 years.
Click to expand...

"Hello, was just trying to play a game today and my PC power cycled itself in the middle of it."

Read this and assumed load dependent. Surely not above rating on the PSU, if functional. Power cycling when running under full load on a questionable PSU? I'd say that's possible.
 
I dont really question the PSU, I just started an AIDA64 and every voltage is a flat line. The PSU so under loaded i doubt its an issue, and its 80 plus Platinum as well. I left out another thing i noticed, I did an IBT once and noticed after 10 minutes the clock speed dropped to 4.4, like the VRMS overheated or something. Also with the voltage set to auto and the CPU at stock it puts vcore at 1.4v under load. I really think the board just sucks. I got a Asus Strix z270 on the way, got a pretty good deal. I might try to RMA this board or something. Tried to buy one local but they only had 8th-9th gen stuff.
 
kirbyrj said:
As a general rule of thumb, CPU's rarely die and it's usually the board's fault if the choice is between the two.
Click to expand...

That's what I thought, until my 7820X died after about a year (I did get a new one from intel RMA) no silly overclocking either, just a bump to the mesh to 3ghz and a 4.6 all core turbo (which I don't know is standard or not, this was with my taichi's standard bios settings).

Then again mine did not boot anymore at all.
 
I didn't see this addressed so just "in case" did you re-apply thermal paste on CPU/GPU and clean the heatsinks ?
Sudden shut down could be caused by thermal or weak supply. After ruling those out you may start worrying about parts.

Since you have onboard video, can you try using it and remove the GPU.

Also, inspect all PSU connections for worn, corrosion, loose, etc.
 
You must log in or register to reply here.
Back
Top