PC Drivers & Software

[jm18190]

Hello all,

I just opened an AMD support ticket but I thought I might get an answer here as well.

I am having an issue with the RaidXpert chipset driver install. Once the drivers are installed, my PC will no longer allow me to power off, restart, or hibernate properly. Windows will shut down (the rebooting or shutdown screen will go black) but my PC remains powered on requiring me to do a hard shutdown to get it to turn off. RaidXpert appears to be working without issue otherwise, and my disk drives are online without fault. Once I restore Windows to a point before the driver install operation returns to normal. Running AMDCleanupUtility does not correct the issue after installing the driver package, and so far the only way to fix the issue is using Windows Restore to go to a previous point. I did not have this issue when I first installed RaidXpert when the PC was built, but I can no longer get back to that working state.

When I hard shut down my PC my motherboard goes into failsafe mode resetting RAM speed to 2400mhz as if there was an overclock or hardware error, but I have confirmed there isn't one. The issue is purely caused by the RaidXpert chipset driver install.

Any help would be appreciated.

Troubleshooting performed:

1.

-Restore windows (reboot/poweroff functionality restored)

-AMDCleanupUtility to clear all drivers

-Install latest chipset driver 18.10.0418

-Reboot

-Install latest amd-nvme-raid-chipset-driver_9.2.0.41

-Reboot (reboot fault begins)

2.

-Restore windows (reboot/poweroff functionality restored)

-AMDCleanupUtility to clear all drivers

-Install latest amd-nvme-raid-chipset-driver_9.2.0.41

-Reboot (reboot fault begins)

3.

-Restore windows (reboot/poweroff functionality restored)

-AMDCleanupUtility to clear all drivers

-Install RaidXpert drivers that I believe were used when system was built (Radeon-software-crimson-raidxpert-relive-17.30-09aug)

-Reboot (reboot fault begins)

System info:

Motherboard - ASUS ROG Crosshair VI Extreme bios v3008

CPU - Ryzen 7 1800x

GPU - Vega 64

PSU - Corsair AX860

RAM - 16GB G.Skill Flare X 3200mhz DDR4 running at 3200mhz and rated timings/power

OS - Windows 10 Pro 64bit v1803 (latest update)

Storage:

Samsung 960 Pro M.2 500gb (boot drive)

2 x Samsung 850 EVO 500gb in RAID0

Seagate FireCuda SSHD 2TB

Thank you,

[bogdi1988]

Hey guys,

So while in the process of optimizing my system for SSD/NVME I ran into the no shutdown issue again. Through the process of elimination I was able to find the root cause of the issue and replicate it consistently!

If you have pagefile set to "NO paging file" in the virtual memory settings, the 2nd reboot after applying the setting, the restart will be broken consistently. As soon as you set system managed page file again and reboot, the feature works again! Something in Windows 10 + AMD RAID drivers + NO pagefile is what causes this issue to occur! As mentioned I tested a couple times and each time reboot was broken after disabling pagefile.

[Matt_AMD]

Interesting findings and would explain why we could never reproduce it as we never disable the pagefile. Unfortuantely you will need to keep this enabled.

[bogdi1988]

Yup! I kept it on. You should please consider this use case in your testing as many tutorials online that cover SSD/NVME optimizations recommend disabling pagefile once you have a good amount of RAM added to the system.

[bogdi1988]

NOTE: disabling hibernation and the Windows write-cache buffer flushing does not seem to impact restarts. I have tested those as well

[frum]

Thanks Bogdi1988

I can confirm that disabling the Pagefile on the AMD ASUS TUF X570 chipset does exactly the same thing, windows shuts down but PC stays on/Retarting windows get black screen PC stays on. I created multiple restore points and lost hours/days trying to fix this issue which is still happening with the AMD_Raid_Software_v9.3.0.38 driver for the X570 chipset. Without me noticing this post I would still be doing this as most people with large amounts of ram and raided SSD's will be disabling the pagefile. My question to you guys at AMD is why have you not fixed this issue yet or at least include a readme that tells people not to disable the pagefile? Also lets hope some other poor soul finds this rather than doing countless windows installs/debugging.

[kushan]

The real issue here isn't the bug with windows shutting down due to the pagefile being disabled, but that people are disabling their pagefile at all.

https://serverfault.com/questions/23621/any-benefit-or-detriment-from-removing-a-pagefile-on-an-8-gb...

Read up on the pros/cons to disabling the pagefile and you'll find numerous tests and reports that even if you have 1TB of RAM on your machine, disabling the pagefile is a bad idea. Not only does it fail to improve performance, it can actually make performance worse - and that's before you talk about the system instability that can result from it (Not just the bug listed here, but other issues).

You should let Windows manage it for you. If you can afford hundreds of GB's of RAM, you can afford the SSD space to back that RAM in a pagefile.

As much as AMD has a habit of neglecting even the simplest of issues (Still can't run Ryzen master on my virtualisation enabled machine....), this is one of those cases where they can be given a pass because disabling the pagefile is considered an unsupported configuration.

[frum]

Already read up on Pagefiles matey. Also the pros and cons. If I want to disable my Pagefile. I should be able to without it stopping windows rebooting or shutting down. Does not happen on my Intel equivilent machine downstairs which is raided. Also just for the record I do not get ANY system instabillity. I respectfully disagree with you, and furthermore opinion is also divided on the matter. Also just for the record I poo poo your idea that AMD can be "given a pass".

[muldep]

Hello guys, sorry to dig up this thread. 

Im having the same problems, but when i put motherboard on AHCI mode, it restarts fine, but using RAID mode, it doesnt shut down or restard like guys have been reported.

I did the array by the bios and after instaling windows/os, so my system OS is another partition.

Any fix?

[U10CU]

This needs to be highlighted. I just ran into this solution after days of effort and looking around. My goodness...!

[U10CU]

I’ve spent days searching for a solution and came across this. I’m new to AMD and the issue needs to be highlighted.  Thanks so much. 

[zizomass]

TLDR: I turned off my desktop power supply while inside the BIOS panel, and the problem disappeared.

This is a problem I've always had since I started using my NVMe in RAID. I have an Asus x399-a and a Threadripper 1950x. I always keep my computer up to date, whether in terms of drivers, firmware or bios, and this problem always keeps coming back randomly without any possible reason! Suddenly, I tell my computer to turn off, and it "turns off," but the disk led keeps on, and the fans go to maximum speed, and there is nothing that can change that.

Every time this problem appeared, I ended up forcing a shutdown by cutting my desktop power, which eventually always ended up corrupting my Windows installation, forcing me to clean and format my arrays (since recovering an NVMe array with a corrupted driver is apparently impossible).

Unfortunately, this problem had returned in the last few weeks (I made a clean Windows install on my machine at the beginning of the year), but this time I decided to search again, and I found this topic! For a second, this gave me a little hope, but I followed all the advice given here, and nothing solved the problem.

I was already accepting that soon my Windows would be corrupted again and that I would have to format my computer again in a few days, but then I had an epiphany! Nobody mentioned here (at least I don't remember anyone saying it) that even with the computer "off," the led that indicates disk reading keeps lit. I imagine that RaidXpert is doing some background process, whatever it is, that prevents the computer from shutting down properly (which explains why my system always gets corrupt after forcing a shutdown by cutting power off).

So, this time I had a new idea: turn off the system when my discs are not being read! Of course!

Whit that in mind, I opened my MB BIOS panel, the read led turned off, and then I hit my desktop power supply switch, turning everything off and hoping that my system would boot properly the other day.

It looks just like a dumb or silly solution for this problem, right? I definitely agree, but here is the thing, I turned my computer back on today (it started normally, and I was glad it didn't corrupt my system) and, to my surprise, when I asked my computer to shut down again, it just shut down as it should and the problem disappeared!

Honestly, I have no idea why this would make any difference at all, but it did! So, if someone is still having this problem, trying this can whort the shoot. Especially since AMD doesn't look to be trying to solve this issue.