Thanks for your answer.
No, it wasn't long enough. Only did 1 hour test, while still adjusting "sliders" etc., to get a rough idea if this certain setting could potentially be stable or not.
When I have a rough idea that I might have a stable OC - I will let it run for 8 hours.
I am using Realbench 2.56 and I am keeping an eye on WHEA errors while testing, among other values ofc.
However, I did not pay any attention to the 12v parameter, so I have no clue how that looked whilst stress testing.
Also, CPU/GPU overclock didnt failed - it passed. The instability in luxmark only occured AFTER enabling XMP.
I have not tested RB whilst downclocking my CPU, but I just found it weird that it passed RB with CPU & GPU OC, but as soon as I enabled XMP it failed the test on the graphic part of the test (luxmark). And by disabling the GPU OC, the CPU OC + XMP passed RB.
After that I tested my GPU OC in TimeSpy, Superposition & Heaven and it passed everytime no problem.
My current VCCIO/SA voltages is at 1.150v each.
So I am not quite sure where the error is.
I assume enabling XMP does not affect the GPU in any way shape or form right? Which could lead to instability in luxmark.
Regardless to my PSU, I am using a Corsair RM750x, 10 months old. I am not sure if I got spare PCIE cables, I will have to check.
What motherboard is this?
What RAM? You didn't really give us any system specs.
This sounds like your system agent voltage is too low...