crosshair VIII extreme + 5950x, strange unstability issue with D.O.C.P enabled

[fcueto]

Hi,

I have the following setup:
- ROG Crosshair VIII extreme
- Ryzen 5950x
- G.SKill ram F4-3600C16D-64GVK (2 x 32GB)

At "stock" ram speed (1333Mhz), everything works fine, the system is rock solid and I can use it for days without crashing.

When I enable D.O.C.P to get 1800Mhz, the system seems stable (I can run memtest and other stress tools for hours without error), but after 2 or 3 days of uptime, often at idle or when watching youtube videos, the system suddenly becomes unstable : I cannot launch any new task, I even cannot call the task manager, all running tasks becomes unresponsive one after another. Ultimately, all I can do is a forced shutdown/reboot.

I have no WHEA errors in my event viewer.

I tried adding some voltage to the SOC with no success (up to 1.17V). But I'm pretty sure the issue is not related to the SOC voltage because as I said, all stress tests run smoothly.

I tried to disable D.O.C.P and manually set the RAM timings and voltage (up to 1.5V instead of the 1.45V required for DOCP), but no success.
I also tried to lower the ram frequency to 1600Mhz but the issue remains.

Now I'm lost because I don't even know where my problem comes from. Motherboard ? processor ? RAM ?

Can someone help me on this ?

Thank you.

[Kelutrel]

If it blocks the task manager, it is most probably the graphics card or the storage, anything particular to note on those ? In particular, if you are using an M.2 storage device, did you remove the film from the thermal pads under the m.2 heatsink on the board ? If you didn't, that may be your storage unit overheating and causing that issue.

Also, did you change the default motherboard BCLK by any chance ? Did you check the temperatures of the storage/graphics card hardware with a tool like HwINFO ?

In any case, if it is one of those, you should find a related critical error event in the Event Viewer under the System or Application logs around the time it happened. Did you check ?

[Ecmslee]

Have you updated the UEFI on your motherboard? The latest version is 0504 as of the time of this post.

[fcueto]

@Ecmslee : I have had the same issue with all uefi : 0211, 0402 and with the latest 0504

@Kelutrel : I do have a M.2 SSD (WDBLACK SN850 1TB). I have removed the film of the thermal pads, and the temperature is normal at around 54°C.

My graphic card is EVGA 3080 ti FTW3 and is not overheating. Plus I have never crashed while gaming, the crashes always occur when the system is pretty idle, or when I'm watching videos.

Your idea of suspecting the M.2 SSD or the graphic card is indeed good, but remember that when I don't use D.O.C.P the system is rock stable.
I don't see how using D.O.C.P will make the SSD or the graphic card overheat ? especially when the issue never happens while stress-testing the system, but rather when it's almost idle.

This is why the title of my post is "strange unstability issue". I have never encoutered such problem and that's why I'm completely lost on this one. And it's a shame because this setup cost me a lot of money.

[Kelutrel]

@Ecmslee : I have had the same issue with all uefi : 0211, 0402 and with the latest 0504

@Kelutrel : I do have a M.2 SSD (WDBLACK SN850 1TB). I have removed the film of the thermal pads, and the temperature is normal at around 54°C.

My graphic card is EVGA 3080 ti FTW3 and is not overheating. Plus I have never crashed while gaming, the crashes always occur when the system is pretty idle, or when I'm watching videos.

Your idea of suspecting the M.2 SSD or the graphic card is indeed good, but remember that when I don't use D.O.C.P the system is rock stable.
I don't see how using D.O.C.P will make the SSD or the graphic card overheat ? especially when the issue never happens while stress-testing the system, but rather when it's almost idle.

This is why the title of my post is "strange unstability issue". I have never encoutered such problem and that's why I'm completely lost on this one. And it's a shame because this setup cost me a lot of money.

@fcueto Actually in your initial message you said that you tried to disable D.O.C.P, and also tried to lower the ram frequency, but the issue was still there. This is why I discarded your ram as the source of the issue.

It may also look like a memory leak that slowly grows up and causes problems after 2-3 days of uptime. You may want to check, from the task manager CPU and Memory performance pages, the number of Handles and the Paged/NonPaged pool after a day or two, as it may be that a driver or an application is somehow continuously allocating resources without freeing them.

[fcueto]

@fcueto Actually in your initial message you said that you tried to disable D.O.C.P, and also tried to lower the ram frequency, but the issue was still there. This is why I discarded your ram as the source of the issue.

It may also look like a memory leak that slowly grows up and causes problems after 2-3 days of uptime. You may want to check, from the task manager CPU and Memory performance pages, the number of Handles and the Paged/NonPaged pool after a day or two, as it may be that a driver or an application is somehow continuously allocating resources without freeing them.

@Kelutrel : at stock (ie ram freq = 1333mhz), the system is stable. At higher frequencies (I only tried 1600 and 1800mhz), the system is unstable.
I don't think the issue is driver related because if it was, I would encounter the issue @stock frequency.

That's why I think it's related to the frequency but what I don't undestand is that all stress tests are passing (I tried aida benchmark, testmem5, Karhu RAM Test and MemTest Deluxe). Actually the system seems perfectly stable until it crashes like I described, often when idling or watching youtube videos.

[D_BRASCO]

Have you tried to skip setting the bios to DOCP and manually enter the DOCP setting in the memory setting. I've read some posts where that worked,

[RedSector73]

Hi,

I have the following setup:
- ROG Crosshair VIII extreme
- Ryzen 5950x
- G.SKill ram F4-3600C16D-64GVK (2 x 32GB)

At "stock" ram speed (1333Mhz), everything works fine, the system is rock solid and I can use it for days without crashing.

When I enable D.O.C.P to get 1800Mhz, the system seems stable (I can run memtest and other stress tools for hours without error), but after 2 or 3 days of uptime, often at idle or when watching youtube videos, the system suddenly becomes unstable : I cannot launch any new task, I even cannot call the task manager, all running tasks becomes unresponsive one after another. Ultimately, all I can do is a forced shutdown/reboot.

I have no WHEA errors in my event viewer.

I tried adding some voltage to the SOC with no success (up to 1.17V). But I'm pretty sure the issue is not related to the SOC voltage because as I said, all stress tests run smoothly.

I tried to disable D.O.C.P and manually set the RAM timings and voltage (up to 1.5V instead of the 1.45V required for DOCP), but no success.
I also tried to lower the ram frequency to 1600Mhz but the issue remains.

Now I'm lost because I don't even know where my problem comes from. Motherboard ? processor ? RAM ?

Can someone help me on this ?

Thank you.

SOC Voltage for this RAM should be much lower, 0.95v, 1.025v, 1.05v but not exceeding 1.1v in manual setting.


I would try disabling c-states in the BIOS and retest DOCP and see if your problem is gone.

[ImpetuousRacer]

I've just come from a Crosshair VIII Hero WiFi and upgraded to a Crosshair VIII Extreme. Setup worked perfect with the Hero. Same EXACT components moved to this board and I can't get it to boot with DOCP enabled. Had to drop from 3600 down to 3466 to get it to post.

Not sure if I have defective board, or maybe BIOS is just immature?

[fcueto]

OK after several weeks of investigations I finally caught the culprit. It was not a hardware problem but a software problem with Asus Armoury Crate. As soon as I use the fan monitoring I got these strange crashes I described.

I have read many forum posts about the exact same crashes with Nzxt Cam but in the end, it's exactly the same with armoury crate. I wonder if the bug could come from the uefi ?