Question D4 Error with GPU?

[doody]

I have a Threadripper system that I've been using for a while now. 3990x with 128GB Ram and 4x 3090's. Ran flawless. My AIO CPU cooler took a crap, didn't know why but the system kept just shutting off every time I started rendering. Finally I found it was the CPU at 95+ degree's. System was probably shutting off for safety. I ordered a new AIO CPU cooler and 2x 4090's for an upgrade. The AIO wasn't going to show up till 3 days after the GPU's arrived, so I put the GPU's in and all 4 GPU's was working great. Finally the AIO showed up so I unseated the CPU to reapply Thermal Grizzly. Once I put the CPU backing and bios told me bios reset. Once I restarted all I get is D4 errors. System shows nothing but black screen and cannot even get to bios to get in.

I enabled Above 4G Decoding, didn't help. I reseated the CPU just in case, that didn't help. Reseated everything else, no luck. Every time I removed 2 GPU's it works again. The minute I put in 3 or 4 GPU's it gives me D4 Error. Am I missing something? I read some say Disable CSM? Does that help? I cannot figure out what the issue is, I'm thinking at this point the motherboard went bad maybe.

Also it's not the PSU, I have two running. Main system and 2x GPU's running on 2050w PSU and secondary 2x GPU's running on 1600w. PSU

Threadripper 3990x
TRX40 AORUS XTREME (rev. 1.1) Motherboard
128GB Ram
2x 2TB NVMe
2x 4090
2x 3090
2050w PSU
1600w PSU

 

[Colif]

D4. PCI resource allocation error. Out of Resources.
TRX40 motherboards have 56 available PCIe lanes. GPU #1 = 16, GPU #2 = 16, GPU #3 = 8, GPU #4 = 8. Total, 48 lanes used.<< you are under this count.
This persons PC has same error with more nvme & 4 GPU - link

I see this error all over the place and one thing is in common. They all have 4 GPU. It doesn't matter which GPU, but it seems to be the one link they all have.
Its a Gigabyte BIOS problem. Not just this board, I see it on workstation boards as well.


I can find this example where the D4 error solved by running off one PSU - they were trying with 4 3090.

Fixed it by switching to single 2000w power supply. No idea why that fixed it, but the problem was completely resolved after that. Been running well for close to a year!

link

 

[doody]

D4. PCI resource allocation error. Out of Resources.

I can find this example where the D4 error solved by running off one PSU - they were trying with 4 3090.



link

Thank you, but I've been running dual PSU this whole time no issue. And your not gonna be able to run a system with 2x 4090 and 2x 3090 with one PSU. That's 1600w in just GPU draw.

 

[Colif]

i edited my post. it appears problem more associated with having 4 GPU.
I can't find an actual reason why but you not only one with 4 GPU and that error on Gigabyte boards.

The PSU thing was just a chance.

I would love to help more but I can't find much more. No clues to solutions.

 

[doody]

i edited my post. it appears problem more associated with having 4 GPU.
I can't find an actual reason why but you not only one with 4 GPU and that error on Gigabyte boards.

The PSU thing was just a chance.

I would love to help more but I can't find much more. No clues to solutions.

I feel like it could be something with the BIOS reset. I mean I had four GPUs running for 2 years straight on this system. Is there anything other than "Above 4G Decoding" that I need to change in the bios?

 

[Colif]

I don't see any mention of anything else apart from that setting.

I would suggest asking on here since the US forums don't seem to work anymore - https://forum.giga-byte.co.uk/

 

[doody]

I don't see any mention of anything else apart from that setting.

I would suggest asking on here since the US forums don't seem to work anymore - https://forum.giga-byte.co.uk/

Thank you, I have already sent a support ticket to Gigabyte, waiting a response. Will try the forums also. This is so frustrating because I had it all working, not sure what changed that. I want to buy 2x more 4090's to replace the 3090's but if it's not gonna work now.....

 

[Colif]

let us know if you get an answer as it might help someone else in future

 

[doody]

let us know if you get an answer as it might help someone else in future

Absolutely, I will report the issue and fix. I know there is a fix because I had it working until I reseated the CPU and Bios reset.

 

[doody]

Okay so I removed the CPU and I found 2 possible 3 bent pins. Gigabyte says that can cause a D4 error with multiple CPU's. I'd have to send the board in to be repaired.