Cafuddled

I had something like this with my HDMI 2.1 cable on my 4k 120Hz LG oled. The cable had a tug and all things seemed fine until it would get random garble on the screen. Sometimes reseating the cable worked, sometimes it did not, same with turning the TV off then on again. Long story short, when my HDMI cable had a hard tug ripping it out, it actually caused the GPU to ever so slightly be poorly seated in the PCI-E slot. Reseating the GPU resolved all my intermittent screen garble.

An interesting one this. I've got a Tuf 3080 on a MSI tomahawk x570 with a 750w EVGA PSU. I've got a 5900, though, but for comparison it should not make much of a difference. If you can't even post, we know that single issue has nothing to do with the OS... unless you mean it does not load windows? For a post issue, I'd look at the seating of the card in the PCI-E slot as I've had issues odd issues form this in the past. Also a good idea to check the seating of the power cables. If you can get into windows again even though it will take a long while, try to see what all your components are running at in task manager or something like OCCT.

Rayteacing uses the GPUs tensor cores which will require more power and raytracing also causes increased CPU load. So increases the bottleneck and power draw when RT is on. Your best bet is to play around with the curve optimizer in MSI Afterburner and undervolt + max out the curve at a frequency you know the heaviest workload would never take the card over 350w. I have a 5900x, a 3080 and a 750w PSU. I do the above to have my GPU run at 300w max at 1980MHz + default memory speed. With a PSU you don't even want to get it close to running at the maximum rated power. Running a PSU near rated power will heavily reduces it's lifespan and increase the possibility of a catastrophic failure, which with a PSU carrys the risk of frying multiple components. Figure out what all the components power will be at maximum combined then add 30%, get what ever PSU that works out to be.

I've seen this a number of times over years now. I think it's something to do with the sleep state or S levels. Sometimes a firmware update will fix it, sometimes a Windows update. CMOS clear can help, or adjusting the S level in the BIOS, however the last suggestion can often mean a Windows reinstall is needed. I think they are talking about remaining slow post wake.

Such a shame, short of taking it to somewhere you can properly troubleshoot at, you're kinda stuck. I'm surprised no computer stores are up for it, they can charge you a small amount of time and then you can buy the problem component from them.

There are still seats even now.

Ooof, maybe carpooling is worth while... mind you for safety reasons I bet they would be against suggesting that.

Few

On lunch guys

Ah, thanks for the update!! I could have sworn they said this week in the last WAN show... Mind you the next WAN show is kinda this week, thanks a million. (Puts phone in the safe were it belongs)

Can we at least get a hint where the release is going to not be, between the forum, YouTube floatplane and reddit its kinda exhausting... I deplore social media, so if it's on Twitter or it's likes I'll miss this release by a country mile. P.s. I'm on vacation in Mexico right now which makes it even harder, I'm so surprised my phones remained in the beach bag so far! P.p.s. I don't think they water down the drinks in this resort... I'm doing this whole thing in hard mode right now, give a vancouverite a break here!

So I did test fortnite and apart from the shader caching stutter I went through a game fine with this 5900x and an 570x msi tomahawk mobo with 3200 cl16 ram. This was after I enabled fTPM again. On my mobo it's off by default and upgrading the BIOS forces it off again, from loads of HP products I work woth, they all do that as well. It used to be if you wanted to have a worth while bitlocker implementation you would enable TPM, however Windows 11 uses it for credentials and encrypted handshakes it looks like. On windows 10 I never enabled it, ever. It only went on because of Windows 11... I'm lucky though, not a shred of stutter. But then I was very very careful to give the system the best mixture of hardware as I knew the exact type of ram made a big impact woth AMD, hence the 3200MHz dual ranked cl16, it's native speed for the bus and being dual ranked it's been shown to improve the CPUs performance by quite a large degree. I wish you just had another working machine without stutter you could drop parts into like I was saying a while back. But short of that, it's going to be a nightmare trying to diagnose this.

Maybe I'm ignorant here, but if you have fTPM disabled in the BIOS, which will work on Windows 10. Does this not remove fTPM as a factor? From what I read back in the day it was mainly a bug with Windows 11, that required TPM to allow the installation. Full disclosure I use Windows 11 on a 5900x and did not see stutter in games before and after the update (don't really play UE4 games), but since I did update the BIOS I've been rolling with fTPM disabled (TPM is typically disabled when updating BIOS) which Windows 11 likes to warn me about.

It's been a little while, any update on rough times when it could be happening, rough times to look for when tickets will go up and/or methods (floatplane exclusive, priority ect)?

It's certainly old-fashioned and crappy, but not unheard of. Think of it like additional resistance to users frivolousley RMA thing's. Personally I feel they should just have better front line support that can rule that out and eat the cost of shipping... but then maybe their accountants have more day than other companies these days.