Octane with Skylake CPUs?

[Scott762]

It's been recently announced there is a bug with Skylake CPUs causing the system to freeze with complex workloads.
http://arstechnica.com/gadgets/2016/01/ ... workloads/

I'm about to upgrade, as I am having a lot of issues with Octane not working. My motherboard and CPU is 4 years old now, wondering if a new board, CPU & DDR4 memory would help.

I have an Asus Sabertooth P67 with an i7 2600K processor, and two Titan X video cards running Windows 10 Pro x64. Latest nVidia drivers - 361.43. Very often Octane either gives an error opening the Viewport, or the Viewport opens and it is just a blank grey with a CUDA error 208. (CUDA error 208 on device 0: The resource was already mapped. -> failed to bind device to current thread device 0: failed to initialize resources).

Once in a rare while I see a video driver crash, and that shuts down both Octane and Poser 2014.


I re-installed Windows 10 cleanly. I had the same problem with my Win 10 system upgraded from Windows 7. Also, the new Star Wars Battlefront apparently doesn't run in SLI for me, wondering if it's my motherboard. It says it runs in SLI, and reports 60 FPS all the time, but sometimes I get some stuttering. That stuttering goes away with SLI disabled, or the other card removed. So there is zero performance improvement with the second card. It's actually faster with one card removed!

[face_off]

Firstly, reinstall Octane Standalone from the installer EXE - that will correctly set your Nvidia time-outs. Then test. If you are still having issues, try rendering with just one card to see if one of your cards if faulty.

Paul

[Scott762]

Firstly, reinstall Octane Standalone from the installer EXE - that will correctly set your Nvidia time-outs. Then test. If you are still having issues, try rendering with just one card to see if one of your cards if faulty.

Paul

Thanks for the reply.

I've already completely re-installed my system from scratch.

I downloaded OctaneRender_2_24_2_win64.exe and OctaneRender_for_Poser_2_24_2_62_Win64.exe.

I also pulled the video cards, trying each one separately twice, as for a for a moment I thought it could be a card. The first time I pulled the second card, it worked every time for 5 or 6 times. Then I swapped cards, all the problems. Back to original primary, still problems like before. Then a few more swaps to be sure, and no change still giving me issues.

I played Battlefront with the other card for a few hours to put it under heavy load, no problems with it.

[Scott762]

The funkyness in Battlefront is what has me wondering if it's my board, from what I read nVidia SLI issues for that game have been fixed a few driver versions ago.

It runs better with SLI disabled. I can run at 4K (3840x2160) at 70% render resolution at 60fps (the engine renders at 70% of 4K, then upscales the output to 4K). Bumping up the resolution to 75% and I would drop below 60 fps at times. But with SLI on, despite the fact that it's a solid 60fps there's some lag, like when framerate drops. Disable SLI and it's silky smooth 60 PFS. After pulling a card, I'm running smooth at 75% of 4K, locked at 60 FPS.

(BTW, all the time I've disabled SLI for rendering).

[face_off]

I'm unsure what error 208 is, and I have seen no other reports of it (and searching these forums shows no other incidents of it) - so it seems to lean towards a motherboard issue. Have you tried rendering the same scenes in Octane Standalone?

Paul

[Scott762]

I just tried it in Octane Standalone. Runs great, I opened 4 of the demo scenes, sometimes closing Octane & loading again and no issues. Tried the different render modes like PMC, direct lighting etc. all good.

I popped in the other Titan X, confirmed that SLI was disabled and opened the same scenes again, all ran great. The few that I recalled the numbers I was also getting 2x the performance too.

[face_off]

Can you export a scene you are getting the 208 error on from the Poser plugin (in ORBX format), then import it into Octane Standalone, and see if it renders pls.

Paul

[Scott762]

Tried that, same results - blows up in Poser, but Renders fine in Octane. Tried two objects/scenes.

I forgot to mention the error is interesting. First time had the error, second time I got the blank grey Viewport. The error:
Traceback (most recent call last):
File "C:\Program Files\Smith Micro\Poser 11\Runtime\Python\addons\OctaneRender for Poser\SetupForm.py", line 641, in EventOpenViewport
File "C:\Program Files\Smith Micro\Poser 11\Runtime\Python\addons\OctaneRender for Poser\UiManager.py", line 5398, in ShowViewportForm
File "C:\Program Files\Smith Micro\Poser Pro 2014\Runtime\Python\addons\OctaneRender for Poser\RenderManager.py", line 516, in StartRender
WindowsError: exception: access violation reading 0x000000002E2FFD20


The directory "C:\Program Files\Smith Micro\Poser 11" doesn't exist. Under Smith Micro, only "Poser Pro 2014" directory exists.

[face_off]

Can you send me a sample scene which causes this error pls?

Paul

[Scott762]

I still had the second card in, trying to render with the Poser plugin. I noticed another error, error 999.
Also when it finally rendered, the second card was idle. Monitoring the GPU usage, at zero.


Started logging on 15.01.16 05:26:10

OctaneRender 2.24.2 (2240001)

CUDA error 999 on device 1: An unknown internal error occurred.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 1: An unknown internal error occurred.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 999 on device 1: An unknown internal error occurred.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 208 on device 1: The resource was already mapped.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources

CUDA error 999 on device 1: An unknown internal error occurred.
-> failed to bind device to current thread
device 1: failed to initialize resources

CUDA error 999 on device 0: An unknown internal error occurred.
-> failed to bind device to current thread
device 0: failed to initialize resources