My own PC woes...

[CrushingAnvil]

I have been building up a pre-built PC I bought in 2020. The only thing original in it is the ASUS B450m+ Gaming motherboard. I'll list the part changes and the order just so the wizards out there have all the info:

1. swapped RTX 2060 and 550w PSU for RTX 3070 and Corsair HX750i (80+ platinum) PSU.
2. swapped Corsair Vengeance LPX 3200Mhz CL16 RAM (2x8GB sticks) for 4x8GB G.Skill Trident Neo Z 3200Mhz CL16.
3. During GPU + PSU swap I recabled with CableMod kit specifically designed for my PSU - didn't seem to have issues with the cables.
4. Swap GPU from RTX 3070 8GB to RTX 3080 12GB ROG Strix OC. Keep 750 80+ platinum PSU.
5. Swap CPU from AMD Ryzen 5 3600 to AMD Ryzen 7 5800x3d (greater power draw but I ran the numbers and I was confident I'd have headroom).
6. Swapped DeepCool Gammax V2 GTE for EKWB AIO cooler. Telemetry doesn't show signs of irregularities. CPU doesn't seem to run hot but my GPU runs extremely hot but it has never seemed to be an issue.
4. Replace stock case fans with Noctua industrial case fans + a Corsair one for exhaust since it's visible and looks nice lol.

So here's the issue:

Occasionally the display connection (DisplayPort or HDMI - it doesn't matter which is used) will drop off and the PC will go into limbo. The screens tend to remain black, although sometimes it reboots and the screens resume normal operation.

I have had many theories about what is going on but none are proven.

1. GPU sag causes connection disturbance on PCI-e lane.
2. CPU getting maxed out (unlikely).
3. Having a large amount of Chrome tabs open while gaming (doesn't seem likely given the amount of RAM and my CPU)
4. Power supply doesn't have enough headroom to entertain CPU and GPU spikes in power draw. I still need to put my PC on the bench and reseat all the PSU cables but I have checked the GPU ones - I'll have to check again.

Sometimes I can game for a whole night without this happening. The PC performs more or less within spec when everything works, and it doesn't show warning signs. It performs amazingly (120fps @ almost max settings in The Division @ 3440x1440 resolution).

Hope you guys have any ideas.

 

[TedEH]

When you say GPU runs hot - how hot are we talking? I have a 3080 that (after re-pasting etc) is able to keep reasonable temperatures, and it's even one of those alienware ones that have notoriously bad coolers on them. This would be the first place I look.
When you say GPU sag - do you mean it's physically sagging? I'd probably fix that even if it isn't the root of the problem.

I think you'll have to go process-of-elimination on something like this. I'd assume the heat or the power, since most games aren't CPU bound, and low memory is more likely to crash your game than your video driver - you should be able to confirm how much is being used pretty easily anyway.

 

[unusual]

If it's not the PSU, it's probably drivers. Go into Event Viewer and look around system logs.

 

[CrushingAnvil]

If it's not the PSU, it's probably drivers. Go into Event Viewer and look around system logs.

Hard to make sense of this tbh with you haha

 

[TedEH]

Digging through the event viewer for hints could be helpful, but I'll again emphasize that you're going to have to narrow it down by some kind of process of elimination - make small changes one at a time to see what effect it has, if any. There's no real reason to rule out any particular piece of hardware, especially if you've been swapping stuff out when this started happening.

If it was purely a driver problem, I'd have expected to see a blue screen, or some other erratic behavior instead of just things going black. If you haven't been keeping those up to date, that's an easy step.

3080 is a really power-hungry card, and 750w is generally the minimum sized power supply recommended for something like that. If you have other high-power-draw components, I wouldn't be shocked if you're butting up against the limits of your power supply. Black screens and sudden reboots don't seem unreasonable to me as a symptom of that. I wouldn't recommend ruling this out unless you've got strong evidence to the contrary.

Also worth checking that your cables are good - I had a screen a while back that used to randomly drop out and it turned out it was just a dud HDMI cable.

 

[CrushingAnvil]

Digging through the event viewer for hints could be helpful, but I'll again emphasize that you're going to have to narrow it down by some kind of process of elimination - make small changes one at a time to see what effect it has, if any. There's no real reason to rule out any particular piece of hardware, especially if you've been swapping stuff out when this started happening.

If it was purely a driver problem, I'd have expected to see a blue screen, or some other erratic behavior instead of just things going black. If you haven't been keeping those up to date, that's an easy step.

3080 is a really power-hungry card, and 750w is generally the minimum sized power supply recommended for something like that. If you have other high-power-draw components, I wouldn't be shocked if you're butting up against the limits of your power supply. Black screens and sudden reboots don't seem unreasonable to me as a symptom of that. I wouldn't recommend ruling this out unless you've got strong evidence to the contrary.

Also worth checking that your cables are good - I had a screen a while back that used to randomly drop out and it turned out it was just a dud HDMI cable.

I suspect you're right that it's the power supply. Might have to suck it up and order a beefier one.

 

[CrushingAnvil]

Just for closure, I have potentially solved this issue by reseating my RAM. Remember, kids - it should click at both ends of the DIMM slots!

 

[unusual]

Good ol' Occam's Razor!