Originally Posted by kl6mk6
I disable my GPUs if I'm just trying to find CPU stability. You can just uncheck them when the stability test starts. Most likely your graphics drivers are crashing and unable to recover. For graphics card testing I use OCCT, Heaven, and the new 3DMark has a stress module apparently.
I was more trying to find out what causes RealBench to give a blank screen (being already 95% inclined to believe it's GPU related problem --- namely drivers) and find a solution
I went from tweaking the installation method to modifying Windows registry but no good result so far.
As I said earlier, I find it to be completely independent
of CPU/GPU & MEMORY frequencies and/or voltages.
You are correct about that though, in older driver versions there would be sometimes a driver crash error, instead of a blank screen.
They "fixed" it now, in 368.39 driver version it's only the blank screen.
I may be a bit obsessing about this issue but having the following stable:
- 12h+ OCCT LinPack 90% Memory & AVX Enabled
- 12h+ AIDA (everything selected but the local disks)
- 8h+ Prime95 Blend+Furmark simultaneously
- 24h+ Prime95 Blend
- 32h+ HCI Memtest
...alongside many others like Memtest86, Heaven, Valley, Kombustor, Cinebench, 3D Mark, Intel XTU etc. (many of them in combination with Prime95 Blend, Small or Large FFT) for the minimum 8h+
... and having random blank screens in RealBench is....nagging me.
In any case, is there any way to disable GPU testing or parts of it in RealBench, maybe in command line mode?
If I could isolate certain GPU tests, I could send my reports to nVidia to maybe aid them in fixing this LuxMark shadow that's been lurking in their drivers.
It could also be a conflict between RealBench and various apps running in background, like hwMonitor, Afterburner etc...it's not really unheard of.
I'll have to try testing on a clean Windows install (drivers only).If I find anything relevant I'll come back on this.Edited by Kargeras - 6/21/16 at 4:18am