Question Black Screen Crash on entire brand new PC

[Sovauthority]

So I bought a brand new PC from scratch:
- GPU: AMD RX 7900XT by Powercolor
- CPU: Intel I7-13700K
- Mobo: Z790 Aorus Elite AX by Gigabyte
- RAM: G.Skill DDR5 2x 16 GB Sticks.
- SSD: Samsung 990 m.2 nvme with a fresh install of Windows 11.

The only thing NOT new is the PSU.
- PSU: Rosewill RBR1000-M Bronze 80 1000W

When I open an intensive game, mostly games with Ray Tracing, my entire computer shuts down. Games and Events such as:
- Clicking "Overclock" on AMD Adrenaline. Clicking "Stress Test" on Adrenaline.
- Halo Infinite (instantly on main menu before I even get to change any settings).
- Control
- Crysis 2 Remastered if RT is on Very High (works when RT is just regular High).
The Outer World Space Cadets: Instant Crash before even the first warning screen is reached, probably due to intensive required performance.
- Witcher 3 works as long as RT is off.
- Pretty much any modern game, even non-RT games that just require high performance. Metro Exodus, Chernobylite, etc..
HOWEVER, RE Engine games run fine.. Resident Evil 4 Remake RT On works flawlessly, same with RE7..But with RE2 Remake, the same reboot issues.

ALL PLAYED ON 1080p, not 4K.

Oh, I also used an LGA 1155 socket Cooler (just bent a little edge for it to fit) for my LGA 1700 socket CPU because NewEgg forgot to include it in my shipping. I have the 1700 brackets now (I have not yet changed them, because I doubt it will be a solution)..
My gaming temps say for Maximum Crysis 2 Remastered or Mass Effect Andromeda (crashes every now and then) or Witcher 3 - 40 - 50 MAX Degrees C for both GPU and CPU.

I tried a bunch of things to fix this. Latest BIOS just released on March 7th, tried all available BIOS (only 4 available).
Turned off PBO option as that seemed to help somebody, not for me.
I tried various AMD drivers as well as switching between "AMD Full" and "Driver Only" install..
Tried to undervolt the card. Both Automatically clicking the button, and manually and tested various percentages (some games work better at 80% voltage, but still crash after a much long playthrough).
The crash is not a black screen crash, but the computer SHUTS DOWN and REBOOTS.. However, RARELY I get a D3D error and the Driver shuts down..

I RMA'd the GPU, and got a brand new one (same kind by Powercolor), and have the same exact issue, so it's not the GPU.

By all numbers, my PSU (since it's the only part that I did NOT changed) should be able to run the card no problem, but I don't have a spare 1,000W PSU to test the process on..

CPU Stress Test and MemTest86 are successful. I do not have any tools to monitor the PSU voltages (and would not understand the numbers either way), but it didn't shut down on the same games when I had my GTX 980 card connected (it would just run on poor performance)..

What are the odds of this being a PSU issue? Is it that obvious of a problem despite the numbers supporting the card? Previous generation games run flawlessly no issue.
I was hoping to save the money for a 4k monitor instead of replacing a PSU, but if any fix solutions can be offered, I would much appreciate it..

 

[Tech Junky]

Put the old card in and test. It might be a transient power spike on the new card when loading things and causing the shutdown.

 

[BTRY B 529th FA BN]

I would go the PSU route and test a new one.

You can monitor the GPU Board Power drawl with Adrenaline overlay. These latest drivers are buggy though

 

[In2Photos]

If you have had this problem with 2 cards I'd say the PSU would be my next test.

 

[BTRY B 529th FA BN]

@In2Photos Yup, there we're no replies when I was posting. That's the more logical route. Wasn't contradicting.

Just to clarify

 

[VirtualLarry]

What's this about having to bend part of the heatsink to fit? LGA1700 sits lower-down in the socket area, I don't think older (stock) heatsinks will cool properly, even if they will mount to the board.

 

[pauldun170]

So I bought a brand new PC from scratch:
- GPU: AMD RX 7900XT by Powercolor
- CPU: Intel I7-13700K
- Mobo: Z790 Aorus Elite AX by Gigabyte
- RAM: G.Skill DDR5 2x 16 GB Sticks.
- SSD: Samsung 990 m.2 nvme with a fresh install of Windows 11.

The only thing NOT new is the PSU.
- PSU: Rosewill RBR1000-M Bronze 80 1000W

When I open an intensive game, mostly games with Ray Tracing, my entire computer shuts down. Games and Events such as:
- Clicking "Overclock" on AMD Adrenaline. Clicking "Stress Test" on Adrenaline.
- Halo Infinite (instantly on main menu before I even get to change any settings).
- Control
- Crysis 2 Remastered if RT is on Very High (works when RT is just regular High).
The Outer World Space Cadets: Instant Crash before even the first warning screen is reached, probably due to intensive required performance.
- Witcher 3 works as long as RT is off.
- Pretty much any modern game, even non-RT games that just require high performance. Metro Exodus, Chernobylite, etc..
HOWEVER, RE Engine games run fine.. Resident Evil 4 Remake RT On works flawlessly, same with RE7..But with RE2 Remake, the same reboot issues.

ALL PLAYED ON 1080p, not 4K.

Oh, I also used an LGA 1155 socket Cooler (just bent a little edge for it to fit) for my LGA 1700 socket CPU because NewEgg forgot to include it in my shipping. I have the 1700 brackets now (I have not yet changed them, because I doubt it will be a solution)..
My gaming temps say for Maximum Crysis 2 Remastered or Mass Effect Andromeda (crashes every now and then) or Witcher 3 - 40 - 50 MAX Degrees C for both GPU and CPU.

I tried a bunch of things to fix this. Latest BIOS just released on March 7th, tried all available BIOS (only 4 available).
Turned off PBO option as that seemed to help somebody, not for me.
I tried various AMD drivers as well as switching between "AMD Full" and "Driver Only" install..
Tried to undervolt the card. Both Automatically clicking the button, and manually and tested various percentages (some games work better at 80% voltage, but still crash after a much long playthrough).
The crash is not a black screen crash, but the computer SHUTS DOWN and REBOOTS.. However, RARELY I get a D3D error and the Driver shuts down..

I RMA'd the GPU, and got a brand new one (same kind by Powercolor), and have the same exact issue, so it's not the GPU.

By all numbers, my PSU (since it's the only part that I did NOT changed) should be able to run the card no problem, but I don't have a spare 1,000W PSU to test the process on..

CPU Stress Test and MemTest86 are successful. I do not have any tools to monitor the PSU voltages (and would not understand the numbers either way), but it didn't shut down on the same games when I had my GTX 980 card connected (it would just run on poor performance)..

What are the odds of this being a PSU issue? Is it that obvious of a problem despite the numbers supporting the card? Previous generation games run flawlessly no issue.
I was hoping to save the money for a 4k monitor instead of replacing a PSU, but if any fix solutions can be offered, I would much appreciate it..

Have the Asrock 7900XT and went through some shennaigans as well.
Turn off adaptive sync in AMD adrenalin.

That was the culprit in my case

If that doesn't work, use 2 independent pcie power cables instead of a shared 1.

 

[VirtualLarry]

PSU: Rosewill RBR1000-M Bronze 80 1000W

What are the odds of this being a PSU issue?

Sadly, being a Rosewill, the odds are high that the problem is the PSU.

Edit: Yes, I am speaking from personal experience with a wide range of models.

 

[Sovauthority]

When I RMA'd the first card I put back my gtx 980 and for good measure my gtx 750Ti. No crashes.

 

[Sovauthority]

Have the Asrock 7900XT and went through some shennaigans as well.
Turn off adaptive sync in AMD adrenalin.

That was the culprit in my case

If that doesn't work, use 2 independent pcie power cables instead of a shared 1.

Yeah the Adaptive Vsync was off by default, I tried turning on and off but yeah my games use in-game Vsync.
I use 1 8 pin cable and 1 6 pin cable to a 6 to 8 pin converter. I'll see if my local store has an 8 pin cable, thank you

 

[Sovauthority]

Sadly, being a Rosewill, the odds are high that the problem is the PSU.

Edit: Yes, I am speaking from personal experience with a wide range of models.

What's this about having to bend part of the heatsink to fit? LGA1700 sits lower-down in the socket area, I don't think older (stock) heatsinks will cool properly, even if they will mount to the board.

I just used sheers to move the screw whole to fit up under the 1700. Ironically I actually just put in the LGA 1700 brackets and surprisingly it cools worse now for an EVO cooler.
I ordered the brackets but they didn't include the mount for the 212 EVO (despite advertised for it). so now ironically it cools worse than my 1155 method (but I had to break the 1155 mount to get the heatsink on, so I can't go back the previous way anyways..

But, I check the temps and it still doesnt spike to 100C instantly for example when launching a game.. it handles a CPU stress test although now instead of a 95C limit it gets to a 107C limit. I'll invest in a new cooler later

 

[Sovauthority]

Sadly, being a Rosewill, the odds are high that the problem is the PSU.

Edit: Yes, I am speaking from personal experience with a wide range of models.

Thank you for your insight.. I talked to GeekSquad at my local best buy (can you believe they wanted to charge me $150 for a stress test lol) but said I can use their PSU for free to test as long as I dont need an agents help. I'll go to that today and report on the results. Thank you.