This one isn't a request for help, it will be more of a detailed log of how I solved my problem (fingers crossed anyway) in the hope it might help others with a similar issue.
I should add that my issue was something specific to having an X570 & Zen 2 pairing while also running PCIe 4.0 devices. If you're not running this setup its unlikely this solution will help you.
The Problem.
After a random period of time (between 10 minutes and a few hours) while specifically playing games (and not all games were affected) the system would hard reboot. By far the worst affected game was SW Battlefront 2 (I'd be lucky to get half an hour before the crash happened) so this is what I concentrated on for testing moving forward.
Testing.
The first thing I did was leave memtest running overnight on Friday. It passed a full pass with zero errors.
On Saturday morning I quickly established that it was only happening in games. I left P95 Small FFTs & Furmark running simultaneously for hours (according to AMD CCC Furmark was running for 6 hours) and the system was totally fine. Temps were completely normal on the CPU & GPU. CPU reached 90c peak & GPU reached 76c peak on the die and 94c peak on the junction. Temps were not the problem.
My next thought was that it was being caused by GPU Hardware Scheduling so I DDUed the AMD HWS Driver and went back to the non HWS but this did nothing.
Next I went into UEFI and disabled EVERYTHING I could think of that might have been causing this, all settings in AI tweaker were set to Auto (or so I thought, more on that next), PBO was set to Disabled, all power limits were set back to 100%, LLC was disabled, heck I even turned off IOMMU & SVT. Rebooted and fired up BF2 and almost instantly it crashed again.
Rebooted back into UEFI and this is were I kinda got lucky, I noticed that while turning everything off in AI Tweaker that I had missed the DOCP setting so my RAM was still OCed however I ignored this because the system had worked for months with DOCP and never had an issue. (Spoiler: I shouldn't have done that).
Anyway I then wasted the rest of Saturday night and a few hours Sunday morning swapping out the GPU, RAM & PSU for spares I had lying around. This also did nothing.
After putting all my original components back I went back to the HWS idea so after doing a full HDD image I proceeded to format Windows 10 2004 and reinstall Windows 10 1909. After installing drivers & Steam/Origin I fired up BF2 and literally as soon as I hit Start Mission the crash happened again. At this point I was starting to get annoyed so after reimaging 2004 back I shut it down and went and watched a movie.
When I came back to it, for reasons unknown, I decided to turn off DOCP. I went back into Windows and got 3 hours of play time in BF2. Went back into UEFI, reenabled DOCP, fired up BF2 and 30 minutes in it went. Progress, its the RAM, right? Wrong.
I noticed that when you turn on DOCP the UEFI warns you that the FClock will be overclocked from the base of 1200Mhz to 1800Mhz (the same effective speed as my 3600Mhz RAM) and that doing this might cause instability with PCIe 4 devices. Then it hit me, the first time I noticed this issue was around a month ago, the day after I installed a Corsair MP600 PCIe 4.0 SSD. At the time I put it down to a random crash and thought nothing of it (since it had never happened previously). Between then and now I've not really played games on the PC at all (I've been playing a lot on my Original Xbox instead) so it stands to reason I've not noticed any crashes.
The Cause and solution.
I did some Googling and discovered, what I think (hope) is the cause. The FClock (or the Infinity Fabric Clock) is controlled by a little SOC built into the CPU. With just my PCIe 4 GPU installed the SOC was fine with OCing to 1800Mhz but as soon as I added the SSD to the bus the SOC could no longer cope with running both the GPU & SSD at 1800Mhz. The solution was as simple as upping the SOC Voltage for 1.1v to 1.15v (according to the article I read it can go up to 1.2v safely on air).
I've been playing BF2 this morning with my system running at its normal UEFI settings, DOCP on, PBO at Level 1, Power Limits at 120% and LLC on Optimised and so far its been rock solid.
Obviously I cannot say its fixed yet as I haven't had the time to test it thoroughly but its played BF2 for over 2 hours with no crash where as normally it goes within 30 minutes. If it goes again I'll make sure to update this thread.
It was a very frustrating but also very fun weekend. It feels good to track down a problem and solve it.
Nothing would impress your significant other more than LTT Cable Ties.
https://www.lttstore.com/products/20-pack-ltt-cable-ties-first-edition
Put them all in a container under the bed? Or go through and get rid of any cables and stuff that you no longer need, like chargers for old devices you no longer have.
My friends were waiting for me to come play Command and Conquer with them because I was busy watching the latest review (For some Logitech mouse) when one of my friends said:
"You're so obsessed with that channel I'm surprised you don't have a Linus hug pillow"
Then I thought, Hey! that's a great idea! (4 teh lulz, of course)
So if anybody has an image that would work for this that's be great!
If I DO get one made I promise I'll post pics!
My friends were waiting for me to come play Command and Conquer with them because I was busy watching the latest review (For some Logitech mouse) when one of my friends said:
"You're so obsessed with that channel I'm surprised you don't have a Linus hug pillow"
Then I thought, Hey! that's a great idea! (4 teh lulz, of course)
So if anybody has an image that would work for this that's be great!
If I DO get one made I promise I'll post pics!
OMG! I love the Z2, I have the Z1 and want a upgrade! But the Z2 is very expensive, and you can't get it on contract in the US.
The Z2 has front speakers (I need) and has the same awesome camera, waterproofness, and body that my Z1 has. Also, it has a screen that is useful, unlike the piece of crap in the Z1!
And Linus, the Small Apps are also found on the Z1. (And the EQ)
Linkolinx reacted to Master Disaster in My 3 day battle with my PC randomly shutting down while gaming and how I solved it..png "Funny")
Linkolinx got a reaction from wasab in Need Portrait image of Linus for Hug Pillow
Linkolinx got a reaction from Sauron in Need Portrait image of Linus for Hug Pillow
.png "Funny"){kind=small}
