OctaneRender™ Standalone 2.25

[grimm]

I was testing Yambo's test scene from the compilation thread and I had Octane 2.25 lock up on me. Everything was fine until I tried to rotate and zoom the scene. Then Octane became unresponsive and one of my CPU cores got pegged to 100% usage. I checked the GPU and it was also at 100% usage with about 3.3Gbs of memory used (GTX 980). After about a minute the GPU usage went to 0%. I waited for about 15 minutes just to make sure that it wasn't compiling again, but Octane never came back and it wasn't rendering either. I ended up having to kill Octane with the -9 option. My memory usage never went up more than 40% (I have 32 Gbs).

Jason

[smicha]

I was testing Yambo's test scene from the compilation thread and I had Octane 2.25 lock up on me. Everything was fine until I tried to rotate and zoom the scene. Then Octane became unresponsive and one of my CPU cores got pegged to 100% usage. I checked the GPU and it was also at 100% usage with about 3.3Gbs of memory used (GTX 980). After about a minute the GPU usage went to 0%. I waited for about 15 minutes just to make sure that it wasn't compiling again, but Octane never came back and it wasn't rendering either. I ended up having to kill Octane with the -9 option. My memory usage never went up more than 40% (I have 32 Gbs).

Jason

Grimm,

That means your gpu failed (you should see red alert in Octane) and the only known to me solution for this is to reboot a computer.

[grimm]

Thanks smicha, but after I killed Octane I'm able to bring it back up and use the 980 with other scenes and no problems. No reboot necessary, and I never did see a red alert.

[smicha]

But did the 'failed' gpu go back to its normal clocks? (MSI AB)

[grimm]

Yes, it appears so, the render speed is as I expect it. Maybe it works differently on Linux?

[smicha]

It's very hard to say what could be the reason.

Just what came to my mind is that 970 with 4GB had true 3.5GB of vram and the rest 0.5 was much slower - there was some kind of scandal with 970 4GB some time ago. I hope your 980 is free from this issue and this was not due to exceeding 3.5GB limit using Yam's scene.

[grimm]

I suspect that it is an issue with Octane, as the GPU went idle so Octane wasn't pushing data to it (or pulling data from it).

[abstrax]

I was testing Yambo's test scene from the compilation thread and I had Octane 2.25 lock up on me. Everything was fine until I tried to rotate and zoom the scene. Then Octane became unresponsive and one of my CPU cores got pegged to 100% usage. I checked the GPU and it was also at 100% usage with about 3.3Gbs of memory used (GTX 980). After about a minute the GPU usage went to 0%. I waited for about 15 minutes just to make sure that it wasn't compiling again, but Octane never came back and it wasn't rendering either. I ended up having to kill Octane with the -9 option. My memory usage never went up more than 40% (I have 32 Gbs).

Jason

Can you reproduce this? If yes, could you send us the scene? Thank you.

[grimm]

I was testing Yambo's test scene from the compilation thread and I had Octane 2.25 lock up on me. Everything was fine until I tried to rotate and zoom the scene. Then Octane became unresponsive and one of my CPU cores got pegged to 100% usage. I checked the GPU and it was also at 100% usage with about 3.3Gbs of memory used (GTX 980). After about a minute the GPU usage went to 0%. I waited for about 15 minutes just to make sure that it wasn't compiling again, but Octane never came back and it wasn't rendering either. I ended up having to kill Octane with the -9 option. My memory usage never went up more than 40% (I have 32 Gbs).

Jason

Can you reproduce this? If yes, could you send us the scene? Thank you.

Thanks Abstrax,

No, sorry, I wasn't able to reproduce this. It must have been a gremlin somewhere. I tried reproducing all the steps I had done before, but no lockup.

Jason

[Inverse404]

Persisting problems with octane render failing to initialize when used from the Octane for Poser Plugin,
when using NVIDIA graphics drivers later than 358.50 !

Even though the problem occurrs when using the Poser plugin I double post this information here, as that was suggested by
the Octane 4 Poser plugin's developer on the plugin's respective thread:
viewtopic.php?f=45&t=40425&p=269426#p269426

I've re-run the testing with a select few drivers that I still had on disk in a strictly controlled procedure.
In the attached file you will find a text file describing my testing procedure, detailed system information,
software version information, and the produced debug log outputs from CUDA log and poser python error log.
Also my minimalistic test scene is included, although it includes nothing besides the ground and Andy.

On my system unfortunately using Octane through the Poser plugin is not possible anymore when drivers later than 358.50 are installed.
It simply fails on opening the viewport with CUDA #999 of doom.
By Paul's request I have rerun this test in strictly reprodrucible procedure and compiled all relevant system/software and log output information
into a document (see attachment).

Please let me know if you can find out anything about how to resolve this problem, as being locked into the already rather outdated driver 358.50 does not work so super well with some other products, and is likely to become even more of a problem as time goes on.