OTOY Forums

wrapthereal66 wrote:Hi ! suddently my octane irp crash ALL the time ! for no reason !

here is the log :
Started logging on 28.08.18 00:11:56

OctaneRender 4.00 RC1 (4000011)

CUDA error 700 on device 0: an illegal memory access was encountered
-> kernel execution failed(kernel25)
CUDA error 700 on device 0: an illegal memory access was encountered
-> failed to launch kernel(kernel34)
device 0: direct light kernel failed

You seem like using RC1. Please try latest build from first page of this topic. And give better information about drivers and system spects. And in which condition it happened. What's steps to reproduce those?

aoktar wrote:

y81r wrote: I have tried several different Nvidia drivers, and removed the old ones before testing a new diver, without any improvements.
It must be a problem of the plugin and not of the standalone core.
I'm not the only one with this problem so I hope you can help us.

This is my setup:

Win10 Pro
Mainboard MSI (MS7A78)
IntelCorei7-7700K 4,20 GHz
32GB RAM
2x GTX1080ti
c4d r19.068
Octane V4-RC2-R1

Thank you for your hard work!
Yassin

Plugin's itself is not having any cuda software. It's built over standalone core. So it cannot be an issue of plugin alone. Maybe some conflicts between C4D opengl and octane core's cuda structure. Try with software opengl mode to have more ideas.

made a test, also with software open GL, CUDA error 700:

Hi, can you fix this problem? (watch video https://drive.google.com/open?id=1XCbff ... Awa3-GQ1NL)
this problem persist for a long years.

Ahmet,
you need to add more ID that it's Toon mat

see variant

Hello Aoktar,
I have the same problem than "stefano-dear"

The render times are really longer in the latest version (4.0 RC2 R1) than the 3.08.1 version.
Here are the screenshots:

v3.08: 0'47 sec v4.0 RC2 R1: 3'46 sec I've just loaded the scene in both versions.
Do you know why the render times are so longer ?

Hello. I was just wondering if anyone has run any tests using the AI denoiser vs something like Neat Video. What I've been doing lately is ramping up the adaptive sampling leaving a fairly noisy render, and then letting Neat Video pick up the pieces. Works pretty good for me. Has anyone done any tests to see which is faster/more effective?

When rendering larger scenes I get constant crashes. No logs from Octane to show because it crashes Cinema 4D completely but I have attached the logs from C4D Bug Reports

Scene is an automotive scene and i am rendering stills from different angles. Tried using the takes system but was almost crashing immediately each time I tried. I can usually render 2-3 in a row before it crashes out.

Total loaded onto VRAM is about 4.0GBs.

Windows 10 1803
i7-6850K
4x1080ti
Cinema 4D R19.068

Let me know if any other information would help. Any timeline on network rendering working?

FxDesign wrote:Hello Aoktar,
I have the same problem than "stefano-dear"

The render times are really longer in the latest version (4.0 RC2 R1) than the 3.08.1 version.
Here are the screenshots:

v3.08: 0'47 sec

4RC2-R1.JPG

v4.0 RC2 R1: 3'46 sec

3.08.1.JPG

I've just loaded the scene in both versions.
Do you know why the render times are so longer ?

It's confirmed that V4 is slower while you have a lot of movable meshes. It's most likely fixed on new SDK.

aoktar wrote:

FxDesign wrote:Hello Aoktar,
I have the same problem than "stefano-dear"

The render times are really longer in the latest version (4.0 RC2 R1) than the 3.08.1 version.
Here are the screenshots:

v3.08: 0'47 sec

4RC2-R1.JPG

v4.0 RC2 R1: 3'46 sec

3.08.1.JPG

I've just loaded the scene in both versions.
Do you know why the render times are so longer ?

It's confirmed that V4 is slower while you have a lot of movable meshes. It's most likely fixed on new SDK.

please look into multiple phoenix grids and itoo forest scatter objects.

Just had to revert to stable version since it became impossible to work on my project. I had really complex blend mat with ~15 4k+ textures and it was freezing for 10-15 sec on node editor every time i change any value. The same project works just fine on stable version.