PC randomly restarting while gaming

Hey there,
I’ve been dealing with these random restarts while gaming—take God of War, for instance. Usually, it happens just once and then stays quiet, though occasionally it won't even bother restarting and just hangs.
i7-11700F cooled by an NZXT x63 AIO
ASUS Prime B460M-A R2.0
Palit RTX 3070 Ti
ASUS ROG Strix 750W PSU
My gut feeling is that the GPU is the culprit, especially since I've swapped it out for an RX 480 8GB for now, but I want to be certain. Is there any chance this could be a power supply issue, or perhaps something on the motherboard struggling to handle the power draw of a beefier card?

Speaking from my own direct experience, I will reiterate this once more: this phenomenon occurs when the system simply fails to execute a specific graphical process, regardless of which individual component is actually at fault. Typically, the culprit is either the GPU, the RAM, or the motherboard.

In most instances, you are looking at a faulty driver for the graphics card or the board, perhaps some actual memory failure on the GPU itself (or again, just a driver issue, who knows), or even a total electrical or signal blackout involving the RAM or the video card. A chip failure is also on the table.

Given that this is happening to you specifically when the chassis heats up, one can reasonably expect that a contact point is failing during thermal expansion. While pinpointing exactly which connection is failing is virtually impossible, cleaning the contacts and pins before reseating the components might help; if that doesn't work, then you're looking at a full component replacement.

As for the PSU, it is a difficult variable to isolate on its own, though the interface between the power supply and the components is certainly a possibility.

Maybe try lubing up those connectors with some WD-40

darkcyclist92 As David Collins6 pointed out:
In my own experience—and I’m speaking purely from what I’ve seen firsthand—this kind of thing typically happens when the system simply hits a wall and can't process a specific graphical task, regardless of which component is actually failing. Usually, you're looking at the graphics card first, then moving on to the RAM, and finally checking the motherboard.

Usually, you're looking at a faulty GPU or motherboard driver, though it could just as easily be a memory failure on the card itself—or perhaps another driver issue, who knows really. It might even come down to a physical connection problem, like a power or signal blackout between the RAM and the graphics card. Of course, there's always the possibility that the chip itself has finally given up the ghost.

It’s a bit grim when you look at it that way—the idea that heat expansion within the chassis is essentially waiting for a single contact point to lose its grip during a thermal cycle. It’s nearly impossible to pinpoint exactly which pin is going to fail or where it’s hiding, but there is a bit of a workaround. Cleaning the contacts and pins thoroughly before seating any components might buy you some time, though if that doesn't do the trick, you're likely looking at a full component replacement.

The power supply unit itself carries quite a bit of weight, which isn't necessarily a bad thing, but I am genuinely concerned about the connection points between the PSU and the rest of the components. It feels like a real dust trap waiting to happen.

Thanks for the reply.

1. Are you using two separate PCIe cables to power your card?
The RTX 3070 is actually quite a beast, so you really ought to be running two individual power cables from the PSU rather than just one daisy-chained one.

2. Is this card brand new, or did you pick it up used?

3. It’s definitely possible that some old drivers left over from an RX 480 might be causing a headache here.
I'd recommend wiping everything clean using Display Driver Uninstaller (DDU) before you go ahead and do a fresh install of the Nvidia drivers.
You should also grab HWiNFO and FurMark; if you run FurMark while keeping an eye on the temps via HWiNFO, you can see exactly what's happening under load.

LP

driftingridge7 said:1. Are you using two separate PCIe cables to power your card?
The RTX 3070 is actually quite a beast, so you really ought to be running two individual power cables from the PSU rather than just one daisy-chained one.

2. Is this card brand new, or did you pick it up used?

3. It’s definitely possible that some old drivers left over from an RX 480 might be causing a headache here.
I'd recommend wiping everything clean using Display Driver Uninstaller (DDU) before you go ahead and do a fresh install of the Nvidia drivers.
You should also grab HWiNFO and FurMark; if you run FurMark while keeping an eye on the temps via HWiNFO, you can see exactly what's happening under load.

LP

Hey, I'm not using two separate ones; I've got them daisy-chained off a single cable. Could that be the culprit?
Also, if I head out tomorrow to pick up a second cable at a local shop, does it need to be from the same manufacturer, or can I just walk into a place like Best Buy and grab whatever they have in stock?

David Collins6 said:Hey, I'm not using two separate ones; I've got them daisy-chained off a single cable. Could that be the culprit?
Also, if I head out tomorrow to pick up a second cable at a local shop, does it need to be from the same manufacturer, or can I just walk into a place like Best Buy and grab whatever they have in stock?

Hey there, good evening! So, just a little heads-up—it’s generally not a great idea to start mixing and matching different cables like that, you know? It can get messy fast. But, since your power supply actually comes with two separate PCIe cables included, why don't you double-check the box? There might be an extra one tucked away in there somewhere!

LP bok

driftingridge7 said:Hey there, good evening! So, just a little heads-up—it’s generally not a great idea to start mixing and matching different cables like that, you know? It can get messy fast. But, since your power supply actually comes with two separate PCIe cables included, why don't you double-check the box? There might be an extra one tucked away in there somewhere!

LP bok

Unfortunately, all my boxes and extra gear are back home in Chicago.
Which means I'll just have to figure out a new setup from scratch.

David Collins6 said:Unfortunately, all my boxes and extra gear are back home in Chicago.
Which means I'll just have to figure out a new setup from scratch.

Maybe just throw a question out here on the forum to see if anyone has a spare cable lying around. If you look closely at the cable itself, there’s usually a code printed right on it—Corsair uses Type 3 and Type 4 cables, and Seasonic uses the BG33 standard, so checking that code is probably your best bet for finding the right match without any guesswork.
Just a heads-up though, that 3070 pulls upwards of 300 watts; it grabs about 75W directly from the motherboard and gets the rest through the PCIe cables. Since a single PCIe cable is typically rated for 150W, if the card starts pulling more than that, your power supply's built-in protection is going to kick in and shut things down.

LP

driftingridge7 said:Maybe just throw a question out here on the forum to see if anyone has a spare cable lying around. If you look closely at the cable itself, there’s usually a code printed right on it—Corsair uses Type 3 and Type 4 cables, and Seasonic uses the BG33 standard, so checking that code is probably your best bet for finding the right match without any guesswork.
Just a heads-up though, that 3070 pulls upwards of 300 watts; it grabs about 75W directly from the motherboard and gets the rest through the PCIe cables. Since a single PCIe cable is typically rated for 150W, if the card starts pulling more than that, your power supply's built-in protection is going to kick in and shut things down.

LP

Thanks a lot, I'll try to track one down from someone.
Lp

David Collins6 said:Thanks a lot, I'll try to track one down from someone.
Lp

In the meantime, you could always grab their Thundermaster software 😵 and just use that to dial back the power target (or the limit) to about 60%.
If you want to keep an eye on how much juice that GPU is actually pulling, you can just run HWiNFO

https://www.palit.com/palit/thundermaster.php?lang=en

driftingridge7 said:In the meantime, you could always grab their Thundermaster software 😵 and just use that to dial back the power target (or the limit) to about 60%.
If you want to keep an eye on how much juice that GPU is actually pulling, you can just run HWiNFO

https://www.palit.com/palit/thundermaster.php?lang=en

Sadly, dropping the power limit didn't do much for me...

David Collins6 said:Sadly, dropping the power limit didn't do much for me...

You shouldn't even be wasting your time on that. To me, the situation is quite elementary:

1. Either you are dealing with a blatant driver bug (which, frankly, isn't news to me regarding Nvidia chipsets—in fact, I haven't owned an Nvidia card in my entire life that didn't cause issues in 3D environments, which is precisely why I stopped buying them for my gaming rigs ages ago), or

2. There is a fundamental issue with the contact points or connections within your chassis. I cannot emphasize enough how critical those physical connections are when you are attempting to build a rock-solid, high-performance machine capable of handling extreme loads; and what is truly absurd is that these connectivity failures most often occur with brand-new components straight out of the box. I have relied on WD-40 for years, and I couldn't care less about anyone's theories regarding its specific purpose. For me, it is nothing short of liquid magic when it comes to boosting PC performance. Alcohol is simply too "dry" and far too weak to properly clean gold-plated contacts and power junctions, including all the Fe-molexe connectors that distribute pure current.

The fact that this card costs roughly $1,100 doesn't change anything; it doesn't make it any more special than any other piece of hardware. A PCB is just a PCB, a chip is just a chip, and solder is just solder.

Given that the shutdown typically occurs only once during the initial warm-up phase, we can hardly point the finger at the PSU for your predicament. One would assume the wattage of your PSU is sufficient, otherwise, these shutdowns would be happening constantly under any significant load. However, we can certainly conclude that somewhere during the thermal ramp-up of the case, there is a momentary loss or drop in power—assuming, again, that it isn't a driver issue. We already know that processing complex 3D meshes in certain multi-layered games can crash cards regardless of how much GDDR they possess.

If your circumstances allow for it and if you are willing to tinker, try sourcing a new, higher-tier brand of power supply (like a Corsair), not because I am suggesting your current one is garbage, but because a fresh unit will likely provide superior contact and thus stabilize the current delivery to your components. That would be the simplest and most straightforward first step toward resolving this mess.

driftingridge7 said:1. Are you using two separate PCIe cables to power your card?
The RTX 3070 is actually quite a beast, so you really ought to be running two individual power cables from the PSU rather than just one daisy-chained one.

2. Is this card brand new, or did you pick it up used?

3. It’s definitely possible that some old drivers left over from an RX 480 might be causing a headache here.
I'd recommend wiping everything clean using Display Driver Uninstaller (DDU) before you go ahead and do a fresh install of the Nvidia drivers.
You should also grab HWiNFO and FurMark; if you run FurMark while keeping an eye on the temps via HWiNFO, you can see exactly what's happening under load.

LP

Just a quick heads-up for anyone else struggling with this: adding that extra power cable actually fixed my issue. If you're seeing similar symptoms, that should do the trick.
Thanks to everyone for the help!

David Collins6 said:Just a quick heads-up for anyone else struggling with this: adding that extra power cable actually fixed my issue. If you're seeing similar symptoms, that should do the trick.
Thanks to everyone for the help!

👍 👍 👍

Unfortunately, I’ve run right back into the same issue again..
I feel like I've tried just about everything under the sun, yet nothing seems to stick.
At this point, my only real move is to have someone test the graphics card for me; does anyone out there near Los Angeles happen to be willing to help me run some tests to see if the GPU is actually the culprit? Since I honestly have no clue where I'd even take it to get it looked at..

Anything showing up in the Event Viewer?

Go check your BIOS and see if PCI Express Active State Power Management is set to On. On.If that’s actually the case, just go ahead and grab it from the drop-down menu. OffMaybe this will help you out.