Get New GPU Task After Failed GPU Task

[Jesse_V]

I've updated the driver and a GPU task is running. Hopefully, I'll know whether it's succeeded soon.

I recommend that you keep an eye on the log. That way if something goes wrong you can pause the GPU slot to prevent many WUs from failing.

[Space45]

It successfully completed a GPU task, so it appears to be working now. Updating the driver appeared to fix everything.

Thanks,

Space45

[Space45]

Unfortunately, after having a few successful rounds of GPU tasks successfully completing, I'm now getting GPU_MEMTEST_ERRORs again. Any suggestions?

Thanks,

Space45

[Jesse_V]

Unfortunately, after having a few successful rounds of GPU tasks successfully completing, I'm now getting GPU_MEMTEST_ERRORs again. Any suggestions

Per the documentation:
http://fah-web.stanford.edu/MemtestG80/ ... readme.txt

If you suspect that your graphics card is having issues (for example, it fails running Folding@home work units), we strongly recommend that you test as large a memory region as is practical, and run thousands of test iterations. In our testing, we have found that even "problematic" cards may only fail sporadically (e.g., once every 50,000 test iterations). Like other stress testing tools, to properly verify stability MemtestG80 should be run for an extended period of time.

It wouldn't hurt to try again and let it run for much longer. If MemtestG80 identifies a problem, then there may be something wrong with the card itself. Its detailed testing has helped me out before.

[Space45]

"Final error count after 50000 iterations over 100 MiB of GPU memory: 0 errors"

[Bleeder]

A few month ago I was also getting the GPU_MEMTEST_ERROR failure. It turned out that it was because I had Remote Desktop enabled and when I actually connected to the machine via RDP the video driver resets and the F@H core dies.

Are you using RDP to connect to the failing machine? If you are the fix is very easy - disable Remote Desktop and reboot the machine then never connect via RDP again.

[Space45]

A few month ago I was also getting the GPU_MEMTEST_ERROR failure. It turned out that it was because I had Remote Desktop enabled and when I actually connected to the machine via RDP the video driver resets and the F@H core dies.

Are you using RDP to connect to the failing machine? If you are the fix is very easy - disable Remote Desktop and reboot the machine then never connect via RDP again.

No, I'm using the computer itself. However, if those or similar services were enabled, would they affect Folding@Home?

[P5-133XL]

Microsoft Remote desktop causes folding failure when you connect. Other alternatives like VNC do not have that issue.

[Bleeder]

No, I'm using the computer itself. However, if those or similar services were enabled, would they affect Folding@Home?

Enabled is fine just don't connect. But if you are not connecting via RDP then this is not the problem unfortunately.

[bruce]

Some screensavers have also been reported as resetting the GPU just like RDT. It can also happen if the computer sleeps.

[Space45]

Some screensavers have also been reported as resetting the GPU just like RDT. It can also happen if the computer sleeps.

So could my computer sleeping due to inactivity cause the GPU tasks to fail?

[mmonnin]

Very well could.

[Space45]

I looked at my power settings and my computer is set to never sleep.

[bruce]

Some screensavers have also been reported as resetting the GPU just like RDT. It can also happen if the computer sleeps.

So could my computer sleeping due to inactivity cause the GPU tasks to fail?

Yes, but not always.

That's one reason why V7.3.6 disables sleeping. An enhancement is currently being discussed in ticket #988

[Jesse_V]

Hmm. As Bruce mentioned earlier (viewtopic.php?f=61&t=23820#p238166) are you overclocking? What temperatures does the GPU get to?
If you don't have temperature monitoring software, there are a number of programs you could use. Common ones include GPU-Z and Speccy, my personal favorite.