ragp13

I did and the reseller couldn't initially reproduce it, but a new 5700XT is stable in my pc and seems to be running much lower fan speeds. Something was off, hard to say what exactly.

I think the cause is the GPU's memory faulting: The DOCP issue was fixed with a newer BIOS, but I kept getting green/black screens or driver crashing. Could be any game, but mostly when either benchmarking Unigine Heaven or TimeSpy Graphics Test 2. These would result in crashes within 5-10 minutes.

I've been experiencing stability issues with my system from the start (month ago), but thought it only happened when I enabled DOCP. However, today I discovered my system crashes (black screen, no reboot) when running Unigine Heaven for about 5-10 minutes. Everything is running at stock speeds (No CPU/GPU/RAM overclock). I've tried undervolting and underclocking the GPU, while running fans at max speed. No effect. Disabled PCI 4.0. No effect Tried drivers 20.7.2 and 20.4.2. No difference CMOS reset of Motherboard. No difference Chipset and BIOS drivers are the latest version. Temperatures are fine (50-60 CPU, 70-75 GPU) Full specs: Sapphire RX 5700XT Pulse AMD Ryzen 5 3600 Asus Prime B550M-A (wifi) Corsair TX550M V2 Crucial Ballistix BL2K8G32C16U4B Any suggestions as to what might be causing this?

Could it have damaged the CPU/GPU or the motherboard itself if it ran at 1.36V for multiple days?

Right, that's what I expected. So why on earth would the BIOS default to this?

So after experiencing instability on a new system with DOCP enabled ( After that, I checked the SOC Voltage in both the Bios and HWINFO64 and the voltage difference when DOCP is enabled for both versions is HUGE! With the newest Bios it's ~1.085V, with the older Bios it's 1.362V! My understandig was that anything above 1.2V is considered detrimental for your hardware? I want to run the older BIOS (with high SOC V) to see if it's more stable than the newer one, but am wondering if it's OK to do so?

They might definitely not be the most expensive versions (e.g., cheapest), but I would expect it to work. Instead, various people with different B550 boards (including STRIX and TUF) have reported similar issues. https://www.reddit.com/r/buildapc/comments/hq2mku/random_reboots_with_asus_prime_b550ma_wifi_and/ https://rog.asus.com/forum/showthread.php?119384-ROG-STRIX-B550-F-GAMING-issues&s=39a427264adf6899eda4f5fe6bd5bbba https://www.reddit.com/r/ASUS/comments/hq0uko/support_intermittent_hard_resets_after_rog_strix/ https://www.reddit.com/r/ASUS/comments/ho00sv/support_please_help_upgraded_tuf_x570_wifi_to/ https://hardforum.com/threads/update-new-asus-x570-bios-update-2407.1998259/page-2 Some have fixed it with bios downgrades and others haven't.

Can you point me towards the topic/post? I'm new here and don't know what PCPP stands for.

Is gaming for several hours enough stress-testing? I'm running 3D mark without crashes as well. It's mostly when going into windows and when closing a game that it crashes. Rarely/never while gaming. I can try and run Prime95 or OCCT's stress test. But even if it crashes, it doesn't tell me what's wrong, just that it's unstable, which I know it is. I can try 1 stick, but I did already swap out the a set of sticks for another. Power limit for GPU is currently stock/default. Can try lowering if you think that it helps, although I would expect it to run fine with stock.

I did use DDU to (re)install older GPU drivers (all in the 20.x range though). However, can DOCP settings and GPU be related? Additionally, all the crashes are hard reboots. The system doesn't freeze or BSOD. It just reboots without a warning. Has also reboot in Safe-mode.

I have cleared the CMOS before (using a screwdriver to jump the connectors), but I can do it again. Memory is in DIMMs 2-4 (or A2 and B2) Tried setting SOC to 1.1 and 1.35V already. Tried higher voltages as well. Haven't tried 2933, but did try 3000 and it crashed. Ran OCCT and Memtest86 and did not provide errors.

Latest BIOS 0805 and latest AMD Chipset drivers Tried it on QVL G.Skill Ripjaws V F4-3200C16D-16GVKB, but also on (non-QVL) Crucial BLS2K8G4D32AESBK Full description of issue here: Really looking for some way to diagnose and solve the issue

I have an Asus Prime B550M-A (wifi) which apparently is one of the lowest tier mobo's. I'm experiencing instability (hard reboots) when I set my DRAM above the stock settings. For example, when I enable the DOCP for QVL G.Skill Ripjaws memory, the system sometimes crashes within a couple of minutes of booting Windows, but sometimes only after several hours (mostly in Windows though, not during gaming). Disable DOCP nets a stable system. Is the B550M's VRM so bad that it can't even get to DOCP rated speeds? I am I wrong to expect that from a motherboard?

Here you go. Just crashed on me within a minute or two of booting with these settings. But weirdly enough quite often the next time it boots after the crash it will be stable for longer. Without anything reverting to default settings, as far as I know.

A screenshot of what exactly in HWinFO? Currently have DOCP/XMP disabled. This is the dump, I think...? DESKTOP-DC1K3BJ.LOG