GPU stress test
reserv888
Posts: 1,143
My computer crashes/black screens often when I make renders or dForce simulations.
I would like to have a stress test that causes the same problem.
I have ran Furmark (and some other tests) but the computer have no problems with those.
It seems to me, from my laymans point of view, that the stresstests doesn't test the mechanisms that are used to make Iray renders or dForce simulations.
Any suggestions of a stresstest that would do the job?
Comments
This has been used successfully when all others have failed;
https://www.ocbase.com/
Thank you PerttiA!
I have now ran the GPU test, twice. Both of them black screened after around 15 minutes. So far so good.
I just cant find any error log where the cause of the crash can be found.
I didn't find anything in the windows log, nor could the program "Who crashed" find anything.
Can't find any analyze/log in OCCT either.
Any advice?
I haven't used it myself.
What do you mean by 'blackscreening'? does the computer shutdown/restart?
I get a black screen. The computer doesn't shut down but I can't do anything, so in the end I have to shut it down hard manually, and then restart.
If the black screens occur while I'm talking to someone on messanger, we can till communicate for a while, so the computer isn't completely dead.
(I have tried "ctrl+shift+WIn+B", but that doesn't help.)
What GPU do you have, and what wattage PSU? When I hooked up my RTX 3090, I had my screen blackout, but it would come back. I had one power cable with two connections to the GPU; it stopped after I put a second power cable from a different power port on my PSU, which stopped my screen blackouts.
GPU: nVidia Geforce RTX 2080 Ti
PSU: 1000W (Seasonic Focus GX 1000)
Have you tried different GPU drivers (older)?
I did hear that the newest driver has a problem. I don't know if they fixed that.
PerttiA and AgitatedRiot
My problems with black screens, several times a day, started three years ago. I have been in contact with nVidia support, ASUS support, Microsoft Support, the retailers support, and DAZ Support. ASUS support said I had to contact the retailer. The retailer claims they tested the card and said it was fine. Everyone seems to think that the cause of the problem lies outside their particular area of responsibility/expertise. So I have been out on a goose chase in search for another hardware problem than the graphics card. During the last two years I have changed:
CPU
Motherboard
RAM
PSU
Monitor and cables
Mouse
Keyboard
I have added a water cooling system to fix the possibility of overheating.
The only things that are the same as three years ago (when the black screens started) are the computer case and the hard drives. And the nVidia graphics card.
I have during these years used a ton of different nVidia drivers, same problem.
I wasn't really going to make this another post about possible solutions. Lots of people have taken too much of their valueable time and tried to help me.
I recently found the program "Who crashed" and while only about 10% of my crashes (black screens) generate something in the Windows log, all of those which actually does, point in the same direction: nvlddmkm.sys.
I have reopened the case with nVidia but they can't find a solution. To eliminate the underlaying suspision that it is DS that causes the black screens, I wanted to recreate the black screens with a stress test tool.
Now i managed to do that but I don't get any error logs to show for it.
Ok, I'm out of ideas at the moment.
If I remember the GPU has two power connectors. Do you have two separate power cables from the PSU running to the GPU?
I am very grateful for all the tips you have given, now and in the past. Thank you.
There are two separate power cables from PSU to the GPU.
NVidia has a tool. CUDA-MEMCHECK | NVIDIA Developer that might help it log errors on the GPU while running DAZ. You will have to do some reading on how to use it.
Thanks for the tip!