Question Pc Reboots under load (Potentially GPU, probably not PSU)

[ReSlice]

Recently I built a new computer with the specs listed below:
CPU: Ryzen 7 7800x3d
GPU: Powercolour Hellhound (Spectral) RX 9070 XT
MOBO: B650 ASROCK Steel Legend Wifi
RAM: CORSAIR VENGEANCE RGB DDR5 RAM 32GB (2x16GB) 6000MHz CL36
OS Drive: Samsung 990 EVO Plus, 2TB NVME evo
Storage drive: Crucial P3 Plus SSD 1TB M.2 NVMe
OS: Windows 11 Pro
CPU cooler: Thermaltake Peerless Assassin ARGB 120 SE
PSU: Corsair RM 850W
Case: Lian Li Lancool 207

Everything seemed fine untill I launched the first demanding game, Jedi survivor (I'm aware the pc performance of this game is horrendous but the problems experienced here are reproducible on other titles). After about 40 minutes, my PC would black screen then reboot, almost in the same manner as if i were to press the power button. I first disabled the reboot on system failure setting and then began to stress test the computer to try and figure out what the issue was.

Running Prime95 for about an hour plus did not cause any problems. Memtest 86 (ran for 5 passes over night) did not detect any issues with my RAM. However, whenever I stress the GPU with fur mark, my computer will reboot after about 30-40 minutes of stress. This leaves me to believe that the problem lies with either my GPU or PSU. As my PSU was the only part of the computer which was not brand new, I thought maybe there was some failing going on there. I bought a new Corsair RM850x to replace my approximately 6 year old RM850 from 2019. The RM850X seems to be rated pretty highly in terms of reliability, so I thought that it be good to test with.

However, running furmark produces the issue once again, even after swapping the PSU. I also am aware that this problem can be caused by overheating; however monitoring temps with HWinfo and AMD adrenaline goes against this IMO. The GPU ,under load, hovers around 60. The CPU can sometimes hit 70+ under load but usually stabilises around 65.

I have attempted to test the problem with a different GPU, resetting drivers and swapping it out for my old RTX 2060. I can run furmark seemingly indefinitely with no problems in this case, however the power draw of my 2060 is significantly lower than my 9070 xt (150w max vs 350w max respectively under load). On account of this, I am not sure this comparison is particularly fair.

I plan to RMA my GPU but I would like to ask for any advice on what could be going on here; I am concerned that my motherboard may be faulty but I am not entirely sure how to test this. The high power draw of the 9070XT could be causing some weirdness with it ( the bios is fully updated fyi).

Thank you guys for any help you can give; this has been a very frustrating experience lol.

 

[Aeacus]

I would like to ask for any advice on what could be going on here

Well, you've already tested the two main culprits;

This leaves me to believe that the problem lies with either my GPU or PSU. As my PSU was the only part of the computer which was not brand new, I thought maybe there was some failing going on there. I bought a new Corsair RM850x to replace my approximately 6 year old RM850 from 2019. The RM850X seems to be rated pretty highly in terms of reliability, so I thought that it be good to test with.

However, running furmark produces the issue once again, even after swapping the PSU.

With this, PSU is ruled out.

I have attempted to test the problem with a different GPU, resetting drivers and swapping it out for my old RTX 2060. I can run furmark seemingly indefinitely with no problems in this case,

While testing the GPU, it indicates GPU issue.

Now, if you want a bit more stability for a short while for your 9070XT, you can undervolt and maybe even underclock it a bit. This usually helps to keep the GPU stable (e.g less temps) but it isn't a long-term fix.

however the power draw of my 2060 is significantly lower than my 9070 xt (150w max vs 350w max respectively under load). On account of this, I am not sure this comparison is particularly fair.

Well, this would only matter IF the PSU would be the issue. But since you replaced your aging PSU with new, good quality unit, it being the PSU issue is slim.

Sure, you may want to hope that it is the PSU issue, because new PSU is FAR cheaper than new 9070XT but symptoms show the 9070XT to be the culprit.

For definitive answer, put 9070XT into 2nd system and stress test it there. If GPU acts up in 2nd system as well, then it's GPU issue.
Or haul it to PC repair shop for diagnostics.

In the end of the day, there is a very good reason why brand new hardware comes with a warranty. Since when no hardware ever fails (be it DOA, instability etc), there would be 0 reason for a warranty.

 

[Roland Of Gilead]

Hey there,

Wait just a minute, before buying or RMA a part. First tell us what bios you are running on the mobo. Have you updated it to the most recent one for your mobo? And if you have, what process did you follow.

Presumably, you did a fresh install of the OS drive, yes? Hopefully it wasn't one from an older system, and which you simply installed and booted up from. That would be a bad idea.

Are all system drivers, including chipset up to date? Same with GPU?

 

[Bob.B]

Recently I built a new computer with the specs listed below:
CPU: Ryzen 7 7800x3d
GPU: Powercolour Hellhound (Spectral) RX 9070 XT
MOBO: B650 ASROCK Steel Legend Wifi
RAM: CORSAIR VENGEANCE RGB DDR5 RAM 32GB (2x16GB) 6000MHz CL36
OS Drive: Samsung 990 EVO Plus, 2TB NVME evo
Storage drive: Crucial P3 Plus SSD 1TB M.2 NVMe
OS: Windows 11 Pro
CPU cooler: Thermaltake Peerless Assassin ARGB 120 SE
PSU: Corsair RM 850W
Case: Lian Li Lancool 207

Everything seemed fine untill I launched the first demanding game, Jedi survivor (I'm aware the pc performance of this game is horrendous but the problems experienced here are reproducible on other titles). After about 40 minutes, my PC would black screen then reboot, almost in the same manner as if i were to press the power button. I first disabled the reboot on system failure setting and then began to stress test the computer to try and figure out what the issue was.

Running Prime95 for about an hour plus did not cause any problems. Memtest 86 (ran for 5 passes over night) did not detect any issues with my RAM. However, whenever I stress the GPU with fur mark, my computer will reboot after about 30-40 minutes of stress. This leaves me to believe that the problem lies with either my GPU or PSU. As my PSU was the only part of the computer which was not brand new, I thought maybe there was some failing going on there. I bought a new Corsair RM850x to replace my approximately 6 year old RM850 from 2019. The RM850X seems to be rated pretty highly in terms of reliability, so I thought that it be good to test with.

However, running furmark produces the issue once again, even after swapping the PSU. I also am aware that this problem can be caused by overheating; however monitoring temps with HWinfo and AMD adrenaline goes against this IMO. The GPU ,under load, hovers around 60. The CPU can sometimes hit 70+ under load but usually stabilises around 65.

I have attempted to test the problem with a different GPU, resetting drivers and swapping it out for my old RTX 2060. I can run furmark seemingly indefinitely with no problems in this case, however the power draw of my 2060 is significantly lower than my 9070 xt (150w max vs 350w max respectively under load). On account of this, I am not sure this comparison is particularly fair.

I plan to RMA my GPU but I would like to ask for any advice on what could be going on here; I am concerned that my motherboard may be faulty but I am not entirely sure how to test this. The high power draw of the 9070XT could be causing some weirdness with it ( the bios is fully updated fyi).

Thank you guys for any help you can give; this has been a very frustrating experience lol.

If you want to play try a different wall socket.
If you used the old psu power cord swap it out.

 

[ReSlice]

Hey there,

Wait just a minute, before buying or RMA a part. First tell us what bios you are running on the mobo. Have you updated it to the most recent one for your mobo? And if you have, what process did you follow.

Presumably, you did a fresh install of the OS drive, yes? Hopefully it wasn't one from an older system, and which you simply installed and booted up from. That would be a bad idea.

Are all system drivers, including chipset up to date? Same with GPU?

Hi, I'm on the latest bios, version 3.30. I updated using bios flashback after a few hours of building the system, as I heard about older asrock bios 'potentially damaging ryzen CPUs (think that was more the case for the 9800x3d but nevertheless i wanted to be safe).
The OS drive was brand new so a fresh install was done upon it. I have also repaired/reinstalled windows as a part of testing. Yes both chipset and GPU drivers were up to date when I was testing everything. I installed and reinstalled (using DDU from safe mode) a few times.

 

[ReSlice]

Well, you've already tested the two main culprits;

With this, PSU is ruled out.


While testing the GPU, it indicates GPU issue.

Now, if you want a bit more stability for a short while for your 9070XT, you can undervolt and maybe even underclock it a bit. This usually helps to keep the GPU stable (e.g less temps) but it isn't a long-term fix.


Well, this would only matter IF the PSU would be the issue. But since you replaced your aging PSU with new, good quality unit, it being the PSU issue is slim.

Sure, you may want to hope that it is the PSU issue, because new PSU is FAR cheaper than new 9070XT but symptoms show the 9070XT to be the culprit.

For definitive answer, put 9070XT into 2nd system and stress test it there. If GPU acts up in 2nd system as well, then it's GPU issue.
Or haul it to PC repair shop for diagnostics.

In the end of the day, there is a very good reason why brand new hardware comes with a warranty. Since when no hardware ever fails (be it DOA, instability etc), there would be 0 reason for a warranty.

Thanks, just wanted a second opinion, as this is only my second time building a computer and first time having any real issues. I had never really done any real testing before so wanted to see if everything I've done and the conclusions I have come to made sense.

 

[Roland Of Gilead]

Hi, I'm on the latest bios, version 3.30. I updated using bios flashback after a few hours of building the system, as I heard about older asrock bios 'potentially damaging ryzen CPUs (think that was more the case for the 9800x3d but nevertheless i wanted to be safe).
The OS drive was brand new so a fresh install was done upon it. I have also repaired/reinstalled windows as a part of testing. Yes both chipset and GPU drivers were up to date when I was testing everything. I installed and reinstalled (using DDU from safe mode) a few times.

When you updated the bios, did you clear CMOS afterwards?

 

[ReSlice]

When you updated the bios, did you clear CMOS afterwards?

No I did not. I am currently on my 2060, I will ddu, reinstall the 9070 xt then clear cmos and test once again.

 

[Roland Of Gilead]

No I did not. I am currently on my 2060, I will ddu, reinstall the 9070 xt then clear cmos and test once again.

Let us know how that goes.

 

[ReSlice]

Let us know how that goes.

Alright, have reinstalled the 9070 and its drivers. This time I installed the chipset and graphics drivers from the asrock website, as there may be the potential that they have specifically customised them to work with their boards (idk thought it was worth a try). I disconnected power and shorted the cmos jumper. Running furmark as we speak, let's hope it all goes well!

 

[ReSlice]

Alright, have reinstalled the 9070 and its drivers. This time I installed the chipset and graphics drivers from the asrock website, as there may be the potential that they have specifically customised them to work with their boards (idk thought it was worth a try). I disconnected power and shorted the cmos jumper. Running furmark as we speak, let's hope it all goes well!

aaaaaaaaaaaand no dice. Reboot at around the 20 minute mark. At this point i deeply suspect the gpu. Secondary suspect probably the mobo. Still confident in the PSU as I've just replaced it; 850W should definitely be enough right, even if the gpu suddenly draws more than 350w in a spike or whatever.

 

[Aeacus]

aaaaaaaaaaaand no dice

Did you try the undervolt and/or underclock as well?

 

[ReSlice]

Did you try the undervolt and/or underclock as well?

I've applied a -20percent power limit on afterburner, testing it out now. Still want to rma the card tho as I don't think I should be forced to do this with the setup I have here

 

[ReSlice]

I've applied a -20percent power limit on afterburner, testing it out now. Still want to rma the card tho as I don't think I should be forced to do this with the setup I have here

same issue, pretty sure the card is just plain defective, waiting on overclockers to respond to my ticket and might ring them up. Thanks for the help you guys have given so far, its appreciated greatly.