Bluescreen, weird NVMe issue - (Ryzen 5950X, Asus Dark Hero VIII, 32GB (3800MHz), NVMe SSD, 3090, Samsung G9 49")

Bug Reporting Hub Crashes (CTDs)
,
1

Do you have any add-ons in your Community folder? If yes, please remove and retest before posting.

No

Are you using Developer Mode or made changes in it?

No

Brief description of the issue:

Since SU5 builds I have not been able to fly at all (or for a couple of minutes, at most).
FS starts normally but freezes in Main menu, or right after FLY.

If Task Manager is running, NVMe usage jumps suddenly to 100% for a couple of seconds, and falls to zero after that. Looks like FS does something that crashes NVMe controller.

After about 30 seconds windows crashes with bluescreen, if FS is installed to the same physical drive as Windows itself.

This may be important: If FS is installed to another drive, Windows does not crash but drive containing FS is unstable and cannot be checked, for example, and event log contains random disk errors indicating that the drive has disapperared.

First this issue occured on Windows 11 (22000.100) but I reinstalled (cleanly) latest Windows 10 to completely new M.2 NVMe drive (with previous drives removed). Exactly the same issue still persists on Win10 and on another NVMe drive, connected to another M.2 slot on the motherboard. I also tried to run win/fs from PCIe M.2 adapter. Not any better.

I have tried to remove Core perfomance boost on CPU, running RAM at standard 2133MHz, even underclocked GPU (Zotac 3090) and GPU memory to make sure this is not a temperature/cooling issue. All components are cool enough. I can run memory tests, CPU and GPU torture tests for hours without any issues and all other games work as expected.

Motherboard has been updated with the latest BIOS and AGESA firmware, and there are no other hardware connected. I also tried to remove all USB devices (except kbd/mouse) without any help.

I have flown with previous FS2020 builds without any issues on the same system.

NVMe drives tested:
Samsung EVO 970 1T
Corsair MP500 512G
Corsair MP600 Pro 2T

Nvidia driver: 471.41, running 5120x1440 display (tried 60, 120 and 240hz)

Motherboard: Asus Dark Hero VIII, latest 3703 BIOS (same issue with 3702 and the one before that, don’t remember the build number)

PSU is Corsain HX1000i, all voltage lines seem to be stable and power usage does not peak before crash - stays around 160-170W.

Anyone else with similar setup?

2

I finally got it working. Have had not even single CTD anymore.

The root cause was faulty ATX power cable (between PSU and motherboard). Yes, really!
I replaced it with another cable and all problems went away.

This was a strange problem. Everything worked perfectly with other games and CPU / GPU burn-in torture tests were fine (even at the same time with 500-600W of total power consumption) and heavy overclocking was also possible. I also ran multiple NVMe benchmark and test utilities and they were not able to find any issues.

Anyway, the faulty cable or connector seemed to cause some kind interference which caused NVMe to stop working for a while - but only in FS - I never experienced similar freezes in other games / programs.

So in case of weird crashes in FS, this is another thing to check :slight_smile:

2 Likes
3

OP - almost same exact setup as you:

5950x/dark hero/3090 ftw3/HX1000i/64gb ddr4 4x16gb/980 pro nvme

Phew…so anyways, I had an unrelated crash, Event ID 41 kernel power errors, check start> event viewer and see what you can find related to the BSOD.

My issue was the AGESA on the USBs, wouldn’t work - after last BIOS update 3702 - fixed the issue temporarily but ALSO I switched from the Corsair HX1000i to Asus THOR 1200W - had a couple crashes after that as well so not sure what fixed it, but gone now. Everything is running stock, though.

I RMA’d the 5950x and the 3090 so the only thing I haven’t RMA’d is the board itself. Memory should be fine, I have 4 DIMMs