6800 xt computer crashing on demanding games. psu problem?

[Thebadgamer98]

Specs below. My pc crashes and reboots when running demanding games at higher settings. Just today I had it crash on Like a Dragon: Ishin and Call of Duty MW2. Checking the event viewer gives no clear indication as far as I can tell, only an event ID 41 which reads "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

I have read conflicting reports elsewhere that it may be the pcie 8pin connectors which leads me to believe the cause of my issue is that my 3x8 pin GPU is powered by 2 8x2 pcie power connectors (that is, two 8 pin connectors which each have an extra 8 pin dangling from them, one of which is plugged into the 3rd slot on the GPU).

I wanted to get this forum's advice before buying this new power supply. Thanks!

Here is my build. pcpartpicker detects no conflicts as I have updated the BIOS for the CPU.

The clarify the pcie situation, here's a photo. (Don't mind the vault boy, he's there to stop GPU sag.)

 

[evernessince]

Specs below. My pc crashes and reboots when running demanding games at higher settings. Just today I had it crash on Like a Dragon: Ishin and Call of Duty MW2. Checking the event viewer gives no clear indication as far as I can tell, only an event ID 41 which reads "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

I have read conflicting reports elsewhere that it may be the pcie 8pin connectors which leads me to believe the cause of my issue is that my 3x8 pin GPU is powered by 2 8x2 pcie power connectors (that is, two 8 pin connectors which each have an extra 8 pin dangling from them, one of which is plugged into the 3rd slot on the GPU).

You should never use 2 8-pin connectors daisy chained on the same cable. Your GPU doesn't know that those 2 8-pins are on the same cable so it might pull the maximum rated safe wattage from each as if each is it's own separate cable. In effect it could be pulling double the maximum wattage the cable is rated for. There's a decent chance that's the cause of your issue.

 

[Vya Domus]

It's probably nothing to do with the PSU or PCIe connectors, if that were the case the system would just shutdown and no reboot.

I suspect memory, run the RAM at 2400mhz, or whatever is the lowest default preset in the BIOS see if that fixes the crashing.

You should never use 2 8-pin connectors daisy chained on the same cable. Your GPU doesn't know that those 2 8-pins are on the same cable so it might pull the maximum rated safe wattage from each as if each is it's own separate cable.

It's fine, if the card has 3 connectors and you use 2 cables it's totally adequate, it's a potential problem only when it's a high power draw card with just 2 PCIe connectors and you use 1 cable with 2 connectors.

 

[Toothless]

I'll bet your PSU can't handle the spikes. Had the same issues with my 850w EVGA til I switched to a better 850w.

Also use THREE power cables, don't chain them.

 

[GerKNG]

daisy chaining is completely fine... it won't change much if you'd run 3 single 8 pins in terms of stability and is in spec. (336W per 8 Pin on the PSU Side.)

But...
Your PSU is low end crap and RDNA2 has insane transients. Replace it.

 

[Thebadgamer98]

Consensus seems to be the PSU, since it's non-modular I plan to upgrade. Will this one be enough?

 

[oxrufiioxo]

Consensus seems to be the PSU, since it's non-modular I plan to upgrade. Will this one be enough?
View attachment 318374

It's pretty risk free with amazon and that is a quality unit if the problem persist you can at least rule out the psu as a problem or try a 1000w unit after.

 

[Vya Domus]

It's fine to daisy chain power connectors, the reason people used to avoid doing that is because in crappy PSUs each cable might have connected to a different rail, so if you used 1 cable and daisy chained them it was possible to go over what that rail could provide, if the PSU has 1 rail, which is basically most of them these days, even with the cheapest ones, it's fine, it all goes to the same rail no matter how they are hooked up.

Consensus seems to be the PSU

Try what I said regardless with the RAM before buying a different PSU, I am telling you this is not usually what a problem with the PSU looks like, if it actually triggered OCP or anything like that the system would have shut down not rebooted, the problem is most likely with the hardware not the PSU.

You can also limit the power limit to 50% and see if that changes anything.

 

[Deleted member 223428]

Consensus seems to be the PSU, since it's non-modular I plan to upgrade. Will this one be enough?
View attachment 318374

Great choice, but I'd probably go HAM and splurge for the 1000w.

 

[evernessince]

It's fine, if the card has 3 connectors and you use 2 cables it's totally adequate, it's a potential problem only when it's a high power draw card with just 2 PCIe connectors and you use 1 cable with 2 connectors.

The 6800 XT is a high power draw card at 302w on average in games. It has transients up to 482w:

I'd call that high power draw. In addition, take a peak at his PSU.

I suspect memory, run the RAM at 2400mhz, or whatever is the lowest default preset in the BIOS see if that fixes the crashing.

Couldn't hurt to try, although I'd recommend MemTest over this. If the system is still crashing at base spec, it doesn't eliminate RAM as a potential source of the issue. The RAM can still cause crashes even at base spec. Hence why it's better just to run MemTest and eliminate it as a potential factor altogether.

 

[Thebadgamer98]

It's fine to daisy chain power connectors, the reason people used to avoid doing that is because in crappy PSUs each cable might have connected to a different rail, so if you used 1 cable and daisy chained them it was possible to go over what that rail could provide, if the PSU has 1 rail, which is basically most of them these days, even with the cheapest ones, it's fine, it all goes to the same rail no matter how they are hooked up.


Try what I said regardless with the RAM before buying a different PSU, I am telling you this is not usually what a problem with the PSU looks like, if it actually triggered OCP or anything like that the system would have shut down not rebooted, the problem is most likely with the hardware not the PSU.

You can also limit the power limit to 50% and see if that changes anything.

Yes, missed your ram comment on first read. Will try it before I buy anything ty

 

[Vya Domus]

The 6800 XT is a high power draw card at 302w on average in games. It has transients up to 482w:

Should still be well under what his current PSU can do, I don't think that's the problem, it was the 6900XT that had crazy 600W+ spikes.

 

[Noci]

I would also consider the fact you have a rather hefty CPU for a low mid-end mainboard equipped with not the best VRM's, that could be a culprit here under heavy loads or at least have some influence..
Can you check your VRM temps with CPU under heavy load? Use HwInfo64 for example.

750W should be able to run this system, unless the PSU has some flaws.

 

[Toothless]

I would also consider the fact you have a rather hefty CPU for a low mid-end mainboard equipped with not the best VRM's, that could be a culprit here under heavy loads or at least have some influence..
Can you check your VRM temps with CPU under heavy load? Use HwInfo64 for example.

750W should be able to rub this system, unless the PSU has some flaws.

Not if the 750w PSU is a junky bronze.

 

[Kapone33]

The 6800 XT is a high power draw card at 302w on average in games. It has transients up to 482w:

View attachment 318381

I'd call that high power draw. In addition, take a peak at his PSU.



Couldn't hurt to try, although I'd recommend MemTest over this. If the system is still crashing at base spec, it doesn't eliminate RAM as a potential source of the issue. The RAM can still cause crashes even at base spec. Hence why it's better just to run MemTest and eliminate it as a potential factor altogether.

After driver updates my 6800XT drew a max of 255 Watts with 275 watt spikes.

 

[evernessince]

After driver updates my 6800XT drew a max of 255 Watts with 275 watt spikes.

You cannot measure transients without special equipment. 255 - 275 is within the normal range for that card. Power consumption will always vary by game, TPU found a max in game wattage usage of 326w for example.

Should still be well under what his current PSU can do, I don't think that's the problem, it was the 6900XT that had crazy 600W+ spikes.

PSU's max wattage was never in question, it's quality and ability to handle transients was.

I would also consider the fact you have a rather hefty CPU for a low mid-end mainboard equipped with not the best VRM's, that could be a culprit here under heavy loads or at least have some influence..
Can you check your VRM temps with CPU under heavy load? Use HwInfo64 for example.

750W should be able to rub this system, unless the PSU has some flaws.

Couldn't hurt to check the rest of the system's temps as well (CPU, GPU, VRM). While each one of those components should throttle before they cause a shutdown like this it's best to find out for sure. I'd recommend that the OP use Hardware Info's logging features to log information while they are gaming.

 

[damric]

According to W1zzard, the 6800XT spikes at 579W.

Sparkle Arc A750 Titan OC Review

Almost one year after our original launch review we revisit the Intel Arc A750. Sparkle's A750 Titan OC is a factory-overclocked custom design with a large triple-slot, triple-fan cooling solution that lower temperatures considerably. Manual overclocking potential is also slightly better than...

www.techpowerup.com

Vengeance RAM always suspect with Ryzen. I would start with that.

 

[RJARRRPCGP]

Consensus seems to be the PSU, since it's non-modular I plan to upgrade. Will this one be enough?
View attachment 318374

It's time to retire that motherboard, too, if you have more than a 6c/12t CPU, pretty much.

 

[GamerGuy]

When I built a 2nd rig based on some spare parts after upgrading some parts for my main rig, I stuck an old PSU I'd had gathering dust in my storeroom. Though old, I was really quite confident that the Enermax MAX REVO 1500W PSU would have no issue running this newly built 2nd rig.

As for PSU, minimum 850W Gold rated PSU from Seasonic/Corsair/etc, but IF I were buy a new PSU now, I'd go for a minimum 1kW Gold PSU.

 

[Lew Zealand]

According to W1zzard, the 6800XT spikes at 579W.

Sparkle Arc A750 Titan OC Review

Almost one year after our original launch review we revisit the Intel Arc A750. Sparkle's A750 Titan OC is a factory-overclocked custom design with a large triple-slot, triple-fan cooling solution that lower temperatures considerably. Manual overclocking potential is also slightly better than...

www.techpowerup.com

Not while gaming, only with Furmark. Most people don't speedrun Furmark.

If you look at the igor'sLAB data Evern linked to, you'll see <1ms spikes to ~470W which is similar to the ~440W of the 3080, so nothing out of the ordinary but still something to consider with a less capable PSU.

 

[Deleted member 234997]

Bad motherboard / CPU combo.
RAM is dodgy.

Ignore the daisy chain stuff, I run my 7900 at 400w through a daisy chained 3 pin with zero issues and the 6900 XT before it.

Possibly PSU...

 

[Vya Domus]

Vengeance RAM always suspect with Ryzen. I would start with that.

Yup.

 

[RJARRRPCGP]

For goodness sakes, at least get an MSI B550 Tomahawk. But here, the ASRock B550 PG Velocita, has been less expensive at Newegg and it has the needed power-delivery capabilities.

 

[VuurVOS]

@Thebadgamer98
VRM of that motherboard is maybe busted. It may official support the processor but a VRM quality is decreading over time.

I had system hangs when I upgraded my 3700X (65W) to a 5900X (125W) on a MSI B450 Tomahawk MAX.
In the bios you can put the CPU in eco mode like 65W TDP. Then the system will be likely stable.

After 3 RMA with that B450 motherboard, MSI decided to refund it. I bought in the mean time a different brand B550 motherboard with a overspec'd VRM. No issues anymore.

 

[HD64G]

I would firstly test with stock RAM settings. This RAM is known to not work well with Ryzen CPUs. If this problems persists, then I agree that the PSU should be the culprit.