Octane 4.00 slave crashing

Sub forum for bug reports

Moderators: ChrisHekman, aoktar

Forum rules
Before posting a bug report, please check the following:
1. That the issue has not already been disclosed
2. That the issue is specific to this plugin, and not Octane in general (Try reproducing it in Standalone)
Bugs related to the Octane Engine itself should be posted into the Standalone Support sub-forum.

All bug reports should include the information below, along with a detailed description of the issue and steps to reproduce it.
A. Operating System, including version (i.e. Win 7, OSX 10.11.2, Ubuntu 14.04, etc.)
B. Graphics Card(s) model (i.e. GTX 580 - 3GB, TITAN, etc.)
C. RAM Capacity (i.e. 6 GB)
D. Nvidia driver version (i.e. 7.50, 7.5.22)
E. OctaneRender Standalone version, if installed (i.e. 2.24.2, 2.23, etc.)
F. OctaneRender plugin version (i.e. v2.25 - 2.21)
G. Host application version, including build number if available (i.e. 3ds Max 2016 Build 18.0)
H. A detailed description of the issue and steps to reproduce it (Include Screenshots or video capture), as well as an example scene if applicable.
I. Copies of the Octane Log window and Console window outputs (full text attached as a file to your post is recommended).

Please note that reports of issues inside existing threads will be ignored/removed, and reports may be closed if the reporter does not respond to subsequent queries in the thread.
Post Reply
User avatar (Default)
boxfx
Licensed Customer
Posts: 277
Joined: Fri Apr 27, 2012 9:13 am

And another for you today :)

We have 2 slaves, one with 10 quadro cards and another with 12 geforce 1060s. The 1060 rig refuses to render anything in octane 4. It is installed and running, the cuda libs are installed, drivers are the latest, but every time we launch a render it starts rendering and then errors out with "CUDA error 2 on device 9: out of memory"

All projects do this, even just a cube.

Octane 3.07 runs fine on this slave.

We have run the DD driver cleaner to completely wipe the system out.

Win 10
Geforce 1060 6gb
8 gigs ram
driver 417.01 (plus tried a couple of older drivers)
extra cuda lib for octane 4 installed
c4d plugin 4.00
octane standalone 4.00
c4d 19.068 / 20.030
Top
User avatar
aoktar
Octane Plugin Developer
Posts: 16066
Joined: Tue Mar 23, 2010 8:28 pm
Location: Türkiye
Contact:
Contact aoktar

Take a look
viewtopic.php?f=24&t=69639
Octane For Cinema 4D developer / 3d generalist

3930k / 16gb / 780ti + 1070/1080 / psu 1600w / numerous hw
Top
User avatar (Default)
boxfx
Licensed Customer
Posts: 277
Joined: Fri Apr 27, 2012 9:13 am

This only seems to be a win 7/8 issue though, all our systems are on win 10. Driver 417.22 does nothing to address the problem.
Top
User avatar
abstrax
OctaneRender Team
Posts: 5510
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

boxfx wrote:And another for you today :)

We have 2 slaves, one with 10 quadro cards and another with 12 geforce 1060s. The 1060 rig refuses to render anything in octane 4. It is installed and running, the cuda libs are installed, drivers are the latest, but every time we launch a render it starts rendering and then errors out with "CUDA error 2 on device 9: out of memory"

All projects do this, even just a cube.

Octane 3.07 runs fine on this slave.

We have run the DD driver cleaner to completely wipe the system out.

Win 10
Geforce 1060 6gb
8 gigs ram
driver 417.01 (plus tried a couple of older drivers)
extra cuda lib for octane 4 installed
c4d plugin 4.00
octane standalone 4.00
c4d 19.068 / 20.030
This is most likely related to the issues Notiusweb has reported and it should be fixed with 4.01, by removing the fast GPU upload which has caused issues on systems with PCIe risers/expanders. If not please let us know.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
Top
User avatar (Default)
boxfx
Licensed Customer
Posts: 277
Joined: Fri Apr 27, 2012 9:13 am

Fixed with 4.0.1 Most likely the DCP lag fix for machines with pcie risers (our direct attached system worked, the pcie riser slave crashed.)
Top
User avatar
Hatsize7
Licensed Customer
Posts: 31
Joined: Wed Dec 14, 2016 6:01 pm
Location: Switzerland

Hi boxfx!
Can you please elaborate on the
boxfx wrote:DCP lag fix for machines with pcie risers
?
Did you do anything special other than updating to 4.01?
I also have a slave which produces Cuda error 2 on heavier scenes (small scenes are OK). It is also a system with 12 cards (GTX1080 and GTX1070ti) with PCI-e risers (H110 Pro BTC+ motherboard). The other slaves with direct PCI-e connection are don't have this issue. I'm on 4.01.1 (not R2) and I still have this problem.
Thanks!

edit: forgot to add, the nvidia driver is the latest 417.35 and running Windows 10 Pro
Win 10Pro / C4D R23.008 / Octane 20.1.5-R4 / Nvidia driver 456.38
Intel Core i7 9800X / 96gb / RTX3090
Top
Post Reply

Return to “Bug Reports”