BSODs with Maximus IX Hero,1TB 970 EVO NvME SSD, & Win 10 Pro 1803. Anyone else?

[Neurisko]

I get blue screens when running RealBench and AIDA64. Errors are: CLOCK_WATCHDOG_TIMEOUT or WHEA_UNCORRECTABLE_ERROR

On Tue 7/10/2018 12:51:34 PM your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\071018-7671-01.dmp
This was probably caused by the following module: hal.dll (hal+0x3F520)
Bugcheck code: 0x124 (0x0, 0xFFFF9F81ABDA5028, 0xBE000000, 0x800400)
Error: WHEA_UNCORRECTABLE_ERROR

On Tue 7/10/2018 11:47:27 AM your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\071018-7406-01.dmp
This was probably caused by the following module: Unknown (0x0000000000000000)
Bugcheck code: 0x101 (0x18, 0x0, 0xFFFFF803C5B8D180, 0x0)
Error: CLOCK_WATCHDOG_TIMEOUT

WhoCrashed and BlueScreenView are not helpful. BlueScreenView just points to hal, ntoskrnl, and pshed.

I've reinstalled Windows 10 Pro three times on the 970. The SSD is in port M.2_2 and SATA ports 5 and 6 are free. It's in PCI X4 mode.

I'm running the same version of Windows on SATA SSD (I've got an oldish 512GB Crucial MX100) works fine. Stress tests run fine, no blue screens at all. It seems like it might be the NvME driver and the Windows update as I've seen reports of problems with the Windows 10 Pro 1803 release and SSDs.

I can post minidumps if that would help. Am I missing something simple? The fact that 1803 works fine with my SATA SSD really points to the new SSD as the source of the problem, but I'd love to find a solution so I don't have to return it and deal with SATA speeds until I can get a new motherboard or they fix the driver issues.

System is:
Maximus IX Hero
7700K at stock speeds
32GB DDR4 @ 3200Mhz (XMP)
Windows 10 Pro
Samsung 1TB 970 EVO
512GB Crucial MX100 SATA3 SSD
2x4TB WD SATA3
EVGA Supernova 650W PSU

Thanks.

[Neurisko]

Well, it's not the NvME SSD. I'm getting the BSODs without the NvME SSD in the system. Back to the drawing board.

[chevell65]

What are you doing to support the 32GB of memory besides just setting it to XMP? Did you set DRAM voltage to 1.35?

Have you tried bumping up the VCCIO or VCCSA/system agent voltage?

Are you using a single kit for this memory or is this two kits combined?

Are you using a dedicated graphics card or are you using the Intel iGPU?

[Neurisko]

Sorry, I had pinned most of my troubleshooting on the SSD, so I left out some details.

- RAM is one kit, 2x16GB matched GSkill sticks. XMP automatically sets RAM voltage to 1.35.
- GPU is 6GB MSI Geforce 1060.

I haven't tried adjusting anything else with the power settings aside from turning off some of the "green" power management settings, mostly because I'm not sure what settings to use. I just want it stable at stock speeds.

[Neurisko]

I was able to run RealBench benchmark for the first time all the way through. Then I ran the stress test for 15 minutes, no problems. Aida64 stress test, no problems. CPU max temp 72C.

Reset to default, then:

* XMP for the RAM (3200Mhz and 1.35v)
* vcore to 1.2
* AVX offset to 2
* DRAM timing mode 2

I'm not sure what fixed it. I'll probably do some more testing to try to narrow it down.

[rolldog]

I'm running a Maximus IX Extreme with a Samsung 970 SSD with 32GB of G.SKILL Trident 4000Mhz memory (I have everything in the BIOS set to defaults & auto/no OC), and I'm getting random freezing in Windows 10 1803. I've reinstalled Windows multiple times, a fresh install deleting my partitions, and it still freezes up. Yesterday, I was sble to use my PC for 3 hours without locking up, which is a record. This has been happening before Windows 1803, so I have no idea what's causing it. It's extremely frustrating as this has been happening since finishing my build.

I read somewhere that many people were having issues with the ASMedia drivers and to use the standard Windows drivers instead, but I tried that yesterday, and it still froze up. It happens at random times, while I'm using it and while I'm not. This is so frustrating, I feel like buying all new components and swapping things out until it's fixed, if it ever does fix it. Because I'm running a custom water loop built with borosilicate glass tubing, taking everything apart to swap components would be a PIA.

[Neurisko]

I tried removing most of my drivers and add-on software, but it didn't help. RealBench wouldn't even run in Safe mode, so Safe Mode was useless to me.

Try the changes I did:

* vcore to 1.2
* AVX offset to 2
* DRAM timing mode 2

My PC is rock solid since I made those BIOS changes, and it's the only thing I did differently.

Oh BTW, I'm also on the most recent BIOS. I foolishly updated my BIOS right before installing the SSD, so I'm not sure if updating the BIOS changed some default settings or the new BIOS is just not as solid. Maybe there were Spectre/Meltdown related fixes in the latest BIOS that made a difference, I don't know.

[chevell65]

It was most likely the DRAM timing change to mode 2 that fixed it.

I say this because I noticed others who were running DRAM mode 1 but after updating to EFI 1803 found it no longer stable. Changing to DRAM timing mode 2 fixed the issue for them, however they were using Z370.

[Neurisko]

So after 4 or 5 days of no problems, I had two BSODs in the past 24 hours. Once yesterday while playing a video game (7 Days to Die) and once today while unzipping a file with 35,000 files in it.

Instead of the watchdog clock error, these two errors are WHEA_UNCORRECTABLEs.

I ran RealBench and Aida64 Extreme benchmarks and stress tests with no errors.

I am baffled. I really really hate random problems like this.

[chevell65]

Have you tried bumping up the VCCIO or VCCSA/system agent voltage? Try setting both to 1.125v for now.