Motion Blur Crashing Scene Tweet

[mikefrisk]

The scene renders fine without MB enabled, but crashes when it loads to the GPU if it is enabled. Any simple way to troubleshoot the item responsible?

[juanjgon]

Hi,

Can you please enable the log file in the plugin options panel and attach the file that you get after the crash? In the log file we can see the crash point to see if the problem is related to the plugin or if it is an Octane issue.

Thanks,
-Juanjo

[mikefrisk]

Attached is the log. Also when updating the scene any streaming video locks up and sometimes crashes, can be quite frustrating.

EDIT: Also it seems it's not necessarily motion blur causing it. Even with MB off my computer completely locks up on scene updates.

[juanjgon]

Thanks for the log. The scene crashes at render time after reporting Octane several CUDA 999 and memory allocation errors. It is hard to say what could be a problem without the scene, but perhaps you could try to enable the out of core feature to see if this helps (anyway I don't think that the problem could be related to the lack of memory unless you are working with very high-resolution images).

Do you have this problem only with this scene?

If you want to pack the scene and send it to me, I could try to reproduce the problem here. You can send the scene by pm or to my email [email protected]

Thanks,
-Juanjo

[mikefrisk]

Juanjo,

Thanks for taking a look.

I am using very high resolution images, however my problem extends beyond this scene. I uninstalled / reinstalled all my video drivers, re-seated all my ram and video cards, and reset my bios. Doing that seemed to fix it, fingers crossed. If I encounter this problem again I fear I may have a bad PCIe slot.

[mikefrisk]

Replying to this thread so i dont have to make a whole new one.

Getting a new error during scene updates.

00:05:14 (0314.68) | ... wait for the getPreviewImage function ...
00:05:14 (0314.71) * OCTANE API MSG: CUDA error 702 on device 3: the launch timed out and was terminated
00:05:14 (0314.71) * OCTANE API MSG: -> failed to pin memory
00:05:14 (0314.71) * OCTANE API MSG: device 3: failed to pin film buffer
00:05:14 (0314.74) * OCTANE API MSG: CUDA error 702 on device 1: the launch timed out and was terminated
00:05:14 (0314.74) * OCTANE API MSG: -> failed to pin memory
00:05:14 (0314.74) * OCTANE API MSG: device 1: failed to pin film buffer

What could be failing here? Gpu? Memory? Motherboard?

[juanjgon]

Hi,

It is hard to say ... if you have two or more GPUs, you could try to enable them one by one to see if the problem is related to any of the GPUs. In any case, try to update your drivers to the latest ones to see if this helps.

If you have this problem with only one scene, you could send it to me to try to reproduce the problem here, but if you have this issue with several scenes, probably the problem is related to your system.

Also be aware that if you are rendering with the same GPU used for the display, in some cases your GPU could not have enough free memory to render the scene due to the memory reserved for the system and apps. If you are working with large texture files, enabling the out of core feature could also help.

Thanks,
-Juanjo

[mikefrisk]

I have 4 GTX 1080ti's, i've tried them indivdually and with different scenes. Still crashes.

What does "failed to pin memory" even mean?

[juanjgon]

I'm not sure. It can be some kind of bug, but it is weird that it happens with different scenes, so this is why the problem sounds to me like a hardware or system issue. What Lightwave, plugin and NVIDIA drivers versions are you using?

If you want to attach one of these scenes I could try to reproduce the problem here. A good test could be also trying to render something using Octane Standalone to see if it has also the same problem. You could export your LightWave scene to an .orbx file (from the IPR, for example) to test it in Standalone.

Thanks,
-Juanjo

[mikefrisk]

I'm using all the latest drivers, octane plugin, and lw versions.

It's such a hard problem to nail down. Right now it's working, it almost seems to happen at random...

I did notice, however, my GPU's will reach 85°C+ under full load. Could they possibly be failing due to overheating? Also sometimes when it crashes my computer will drop one of the GPU's effectively only registering 3 out of 4, but the dropped GPU will still work for display (the monitor connected to it still works).

It's so frustrating this is a brand new build with all the latest and greatest.

Edit: Forgot to mention I have not yet tried it in standalone. I will at some point, also I will package the scene and attach it if the problem continues.