Gigabyte H370 HD3 stability issues

[CamaroSS396]

So recently I've decided to upgrade to a non-overclocking rig, hoping for a cheap, hassle-free, all-auto, no worries setup.

I upgraded from Gigabyte Z170 HD3P + i5 6600k to Gigabyte H370 HD3 + i7 9700F.
Turns out I was wrong, since I started getting absolutely random abrupt reboots under no load, as if I pressed the reset button, no BSOD etc. I tested the hardware, disks, PSU (new Thermaltake Bronze 750W) and RAM (two Kingston HyperX @2666) are ok, the rig survived every stress test.

Still, it rebooted when simply browsing or leaving the PC idle, so I thought the CPU lacked power at some point. Turning off C-States did no good, as did changing PLL calibration.

What did help was pumping Vcore to 1.3V. The system became stable, but it produced too much heat and under load the CPU clock dropped to ~3 Ghz.

Now I set the voltage back to Auto, but disabled other settings, Voltage Optimization, Energy Efficient Turbo and Adaptive BCLK voltage. Works fine so far, about 24 hours.

So which piece of hardware is to blame here? Did anyone get this same configuration to work?

 

[EarthDog]

Is the bios for the board updated to the latest version?

Does the event viewer show errors at the reboot times?

 

[CamaroSS396]

Sure, the mobo came with the latest F14 BIOS.
No BSODs, no bugchecks, no events aside from "The system has rebooted without cleanly shutting down first".
The system keeps rolling, so it seems one of "Voltage Optimization, Energy Efficient Turbo and Adaptive BCLK voltage" options are responsible.

 

[trents]

Are you using a discrete video card? What cooler are you using? I take it you are using the OEM boxed cooler since 1.3 volts causes overheating and throttling.

 

[CamaroSS396]

Yes, a discrete GPU. Clearly not an overheating issue, since I use a tower cooler, and reboots happen when CPU is ~30 deg C.
1.3 volts does cause TDP throttling but it also takes reboots away

 

[CamaroSS396]

Aaaand rebooted again, this time while locked when I was away. Also this time it rebooted into BIOS with "optimized defaults". Guess I'm going back to 1.3v solution.

 

[CamaroSS396]

I've noticed something very interesting in sensors data. Is 0.7v a possible Vcore?

 

[trents]

Yes, it is a possible vcore at idle. When running stress tests I always refresh the view after the test has been started for this reason.

Do you need a full 1.3 volts to eliminate the restarts?

 

[CamaroSS396]

1,280 produced one reboot in a day, 1,295 works fine so far.

 

[trents]

Still, something's wrong there. There must be another setting in bios that needs tweaking that doesn't require to run the whole CPU at that high a voltage. What about bumping up System Agent a bit? It's the memory controller component. How much RAM do you have installed? You might also try giving the RAM voltage itself a small voltage bump.

 

[CamaroSS396]

Thank you for suggestion, but that didn't work out.

Then I stumbled upon this thread
https://www.hwinfo.com/forum/threads/pc-random-shutoff-caused-by-hwinfo64.5475/

Indeed, running HWINFO64 increased the chance of reboot greatly in my case too. So I uninstalled it completely.
After this, I could run a CPU @1.2V with no issues EXCEPT one - a guaranteed reboot ~5 minutes after I lock the PC. So I went searching what else could be messing with this ITE chip sensors. I've suspected two quite old drivers/services named gdrv.sys and gdrv2.sys coming from GIGA-BYTE (with hyphen).

Delete both with sc delete, and no issues for 5 hours so far, lock/sleep work OK too. Will post if it finally worked. Gues there was some signaling war going on

 

[trents]

Found this: https://www.reviversoft.com/processes/gdrv.sys

Do you have Gigabyte Tools installed?

 

[CamaroSS396]

Yes, I used to. Sadly, the reboots go on.

 

[CamaroSS396]

So far, the most stable setting in Vcore 1.3 in BIOS (mobo doesn't allow to set it higher even 1.295 crashes); Load Line Calibration: Turbo; Voltage Optimization, Energy Efficient Turbo and Adaptive BCLK voltage disabled.

No crashes with this config, so I guess I'll stay at that. Pumped the TDP in XTU to 100W, stress tests show the cooler can take it.

 

[trents]

I would consider a fresh install of Windows at this point.

 

[CamaroSS396]

I doubt that would help, because without 1.3v it managed to reboot even from an offline chkdsk.
But I sure will consider that

 

[trents]

I also doubt a fresh install of Windows would fix your problem but at this point. Is your Windows power plan set to high performance?

You might also try running system file checker and here is the right way to do it in Windows 10: https://support.microsoft.com/en-us/help/4026529/windows-10-using-system-file-checker

 

[CamaroSS396]

Yup, I know about that. High performance is set, and system integrity OK. No problems back on 1.3v so far, I even installed the May feature update without any problem. So my guess is a *partially* faulty mobo, that is unable to reliably do *something*.

 

[trents]

I think you are correct about the motherboard. Are you considering an RMA?

 

[CamaroSS396]

Nope, it works Don't want to waste any more time and sure would hate to downgrade.