[SOLVED] New PC keeps crashing HELP

[spencer.-]

Hi

I built my first computer a few months ago.

Specs:
Motherboard z370 gaming 3
Cpu i7 8700k
Gpu vega 56 gigabyte oc 8gb
Ram 16 gb one stick
Psu evga gq 650w
Cooler deepcool assasin ii
Os windows 10

As soon as i finished building the pc the problems started.

Pc crashes after 20/30 minutes to an hour and a half of running heaven benchmark. Most times screen goes black and mouse and keyboard too. Gpu fans ramp up to 100% and i have to keep the power button pressed for a few seconds to restart the pc.

After lots of troubleshooting i narrowed it down to a gpu related isue.

The computer works flawlesly without the gpu installed. Also i got the same problem with the gpu installed in another pc.

I also tested with the original psu that i intended to use for the build wich was of less quality.

I was sure the gpu had some kind of isue.

It took it to RMA. They alegedly tested it twice for more than 3 hours and said that it didnt show any problem.

I recived the card again and re tested and i keep having crashes almost every half hour.

I installed the latest bios for my motherboard. I used DDU to uninstall amd and intel drivers and reinstalled the radeon drivers. I tryied messing with the gpu settings, undervolt, power limit, underclock.

I really dont know what else to do.

Do you have any ideas?

I run all tests with hw monitor and there is no thermal isue. Cpu arround 60 on a hot day and gpu between 60 and 75

Thank you so much.

 

[electro_neanderthal]

If you're using Windows, did you make sure to disable the automatic driver installer? That has caused havoc for me before.

This next suggestion is... well, to be frank, I don't think it will fix anything. However, in the one in a million shot this happens to be the case, I'll just say it: Have you tried reinstalling your benchmarking software? I mean, a corrupted file could cause something off during the testing.

Though, really, I suspect that the RMA department messed up somewhere. Just my 2 cents.

 

[spencer.-]

Thanks. I tried both things but problem persists. Maybe its a problem with the motherboard and not the gpu.

 

[spencer.-]

Event viewer shows kernel power 41. Ive allready tried all the fixes i found online for this error but still continue to crash

 

[Ravi Sankar]

It seems your GPU heat management is not proper... to confirm this, you please run Furmark GPU burner and share the graph how steeply temperature is raising? and where it is saturating?

 

[spencer.-]

Hi. Thanks for your help. Here is a link to the furmark screenshot. It seems normal. https://ibb.co/QcbQYSR

 

[Ravi Sankar]

If GPU and CPU temperatures are in safe range and still crashing computer, then there could be chances that sub-components/OS/Software/Virus problems. Now we can look at the module wise, so, please look the event details to find any problem with whatever components you suspect. You can navigate to device manager and check (reference image attached)
https://imgur.com/a/mS4GiDa

If you still could not figure out the root cause, then, I doubt on OS now, I suggest you to check the same Furmark stress test on some other windows like Windows 7

Before Running...
1. You please post idle GPU and CPU load percentage and temperatures, so that we can make sure that there is no virus kind of things
2. Make sure that heat sinks of CPU and GPU are free from dust. (Dust cause airflow obstruction that discourages cooling)
3. Make sure that CPU/GPU fan wings are free from sticky dust, if this is present, then it will drastically reduces the fan RPM

Good Luck

 

[spencer.-]

Yesterday i updated windows and gpu drivers again. For the first time heaven run all night for 6 hours. I was using the x4 pcie because i was trying to rule out a problem with the x16 slot.

So this morning i tried heaven again in x16 slot and crashed. I re tested on x4 slot again and crash. Its so wired...

 

[spencer.-]

Hi

Update. I did a fresh install of windows on another hdd and it worked perfect.

I guess its a problem with the os or with a faulty ssd.

I will try installing windows again. I guess if it still crashes then its the ssd.

 

[spencer.-]

Find the problem. Did a clean install of windows 10 and let windows install drivers on its own. Everything works great until i update the vega 56 drivers. It dosent matter wich version it is. I guess it makes some kind of problem with another driver when re installing. Now i dont know how to fix this isue but at least my sistem si stable only with older gpu driver.

 

[electro_neanderthal]

Find the problem. Did a clean install of windows 10 and let windows install drivers on its own. Everything works great until i update the vega 56 drivers. It dosent matter wich version it is. I guess it makes some kind of problem with another driver when re installing. Now i dont know how to fix this isue but at least my sistem si stable only with older gpu driver.

Glad you found a solution, but it's too bad it's not the optimal solution. You might want to let the folks at AMD know so they can possibly fix it, since it sounds like it's probably hampering more people's systems; they might just not know it yet. I believe www.amd.com/report is the place to do it. And with any luck, you'll see a fix within the year that will let you use the most up-to-date drivers.

Good luck though, and have fun!