|
May 03, 2017, 05:32:09 PM |
|
Hi all! I built 4 rigs, but issues on the 7 GPU RX570 Rig. Constant BSOD Loop when attaching more than 5 Cards to rig.
Here is my specs and issue;
MSI Z97 Gaming 5 - Tested on Gen 1, 2, 3 and Auto. 7x MSI RX 570 4GB Gaming X - all on powered USB risers. Windows 10 Professional 1250W Platinum Rated PSU 120GB SSD AMD 17.4.x (tested 3 and non-wqhl 4)
All my other rigs are the same, except for the 570 GPUs.
The issue is that if I attach 1-5 GPUs to this rig, it will mine and mine well. Once I attach 6 (and/or 7) GPU's, it will boot into windows, recognise the 6th card, declare a BSOD (Tried to execute noexecute memory) and loop. If I detach the 6th and 7th GPU, it will boot into windows but Once I try to start mining in (and do other things), I get a BAD IMAGE error for certain DLL files.
Here is what I've tried;
BIOS Update Different Gen and PCI Bus Clock Speed combos (all combos) Reinstall Windows cleanly (required once the bad image errors come up) Swap Risers, USB Cables, Cards in different slots. Set higher Virtual Memory Tested all drivers that my card uses ( was told I was obligated to use 17.x.x) Flashed the BIOS to a modded one, then tested that, didn't work, flashed back to stock, didnt work.
Here are my theories;
It has to be the new drivers. I loaded SimpleminingOS (I hate linux and its slow hash), all 7 card recognise, and begin to mine. If I load Windows with 7 Cards attached to the board, but with no drivers installed, Windows will recognise all 7 GPUs under Basic Display Adaptor (with only the GPU with Screen Attached showing as active).
I also got this same BSOD loop error when incorrectly trying to load 16.12.x Drivers with my RX470's and not correctly patching the Driver Signer. That always led to the same Attempt Execute NoExecute Memory error.
I am extremely confident it is a Software issue, and I am confident that it is a driver issue.
HELP!
|