OctaneRender™ 2018.1 RC4

[coilbook]

Seriously what is going on. Both 2080 Tis give me error with a brand new scene but titan x pascal works fine . Everything was working fine 4 hours ago. Drivers were reinstalled with driver remover. Did both 2080 TIs fail? Everything was working fine for a month and I never had this problem. using the same 717.71 drivers please see the video attached

[paride4331]

Hi coilbook,
as I can see in your clip, out-of-core is not checked; Titan XP has 12GB on board, 2080TI only 11GB. Could you try using out-of-core option?
Regards
Paride

[PolderAnimation]

Can we have the option for ABC an FBX file to loop, and to have a time offset. This is very handy when using animated ABC files in a large project where sometimes the frame ranges are in high numbers and the looping abc file is only 400 frames long. Hope you guys can consider this.
Cheers

[Notiusweb]

Can we have the option for ABC an FBX file to loop, and to have a time offset. This is very handy when using animated ABC files in a large project where sometimes the frame ranges are in high numbers and the looping abc file is only 400 frames long. Hope you guys can consider this.
Cheers

Yes, i.e., if we import in a Geometry as a 'Scene':
1) Geometry 'Scene-1' (ABC/FBX) - 1,200 frames
2) Geometry 'Scene-2' (ABC/FBX) - 400 frames

...Scene-2 will stop animation after frame 400, whereas Scene-1 will keep going until frame 1,200.

Would be great to have the ability to, in the wrench-icon settings, activate a loop for the animation of an imported geometry scene so that it animates until the end of the greatest imported time frame, such that Scene-1 would play 1x, while Scene-2 @400 frames would play 3x (1,200 frames)

And, what would be even more awesome, would be to be able in settings to set it up so that it loops up to a certain point.
Let's say I want Scene-2 to animate 2x, up to frame 800, and have Scene-1 keep going until frame 1,200. I could set the frame loop end-point for Scene-2 of 800.

@PA- Great idea!
@Otoy - DO THIS

[coilbook]

Hi coilbook,
as I can see in your clip, out-of-core is not checked; Titan XP has 12GB on board, 2080TI only 11GB. Could you try using out-of-core option?
Regards
Paride

Thanks Paride.
We tried both ways. We removed these 2 2080tis and installed another one and it works. Then we reinstalled 2 old 2080Tis They worked at first then quit. Now we just have one old 2080Ti. We also had intel 10Gb x540 T2 switch between 2 2080TIs and that thing gets hot. I wonder if 2080TI vram is more sensitive to heat
I read some articles and they all point to deteriorating vram on 2080TI made by micron. I wonder if heat kills these cards or makes them unstable but at the same time both cards are hybrid Evga 2080Ti FTW3


UPDATE:


Never mind
Still getting these errors. Does it mean bad cards?

CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to wait for event
device 0: direct light kernel failed
denoiserThread0 : Initialization failed. Restart required
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to destroy CUDA event
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to destroy CUDA event
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device array
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated

[Notiusweb]

Otoy - if I F around with displacement too much and too vigorously I can crash Octane, even crash my PC to blue screen with a Win 10 warning about Page Fault in non-paged area blah blah blah....

Is there any setting, or recommended rule-of-thumb set-up, in Octane that you know, or users have come to discover, that would tend to make it more stable as far as being able to go crazy with the displacement sliders? And by crazy I mean whipping the sliders around randomly because I don't feel like typing in the numbers because I'm having too much fun moving the sliders and seeing crazy sh!t. (ie I drag it from .01 to 5.0, kind of haphazardly and accidentally.)

Thx!

[bagel]

Otoy - if I F around with displacement too much and too vigorously I can crash Octane, even crash my PC to blue screen with a Win 10 warning about Page Fault in non-paged area blah blah blah....

Is there any setting, or recommended rule-of-thumb set-up, in Octane that you know, or users have come to discover, that would tend to make it more stable as far as being able to go crazy with the displacement sliders? And by crazy I mean whipping the sliders around randomly because I don't feel like typing in the numbers because I'm having too much fun moving the sliders and seeing crazy sh!t. (ie I drag it from .01 to 5.0, kind of haphazardly and accidentally.)

Thx!

My guess there isn't too much that OTOY can do to alleviate this. Sounds like you're having a buffer overrun of some type. They want the sliders to be immediately responsive so they send the updated data to the render engine as soon as the slider is moved. If you send too much data, it can crash because once that input buffer fills up, there's no place for those inputs to go. They could make it so that the slider doesn't send it's value to Octane until you release it but it would greatly compromise one of the best features of Octane which is immediate feedback. My advice? Just slow down a little bit.

[haze]

now these new errors appear without any reasons. these scenes were rendered for months and all the sudden all these scenes do this. Nothing was changed Even isolating one object this scene does this


I was rendering with these drivers for a long time and all the sudden i cannot even open octaneviewport without an error .

denoiserThread0 : Initialization failed. Restart required
CUDA error 702 on device 0: the launch timed out and was terminated
-> kernel synchronization failed
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to load symbol data to the device(const11)
device 0: failed to upload imager params
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to unregister pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to unregister pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to destroy CUDA event
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to destroy CUDA event
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated

Are you able to PM the scene?

[coilbook]

now these new errors appear without any reasons. these scenes were rendered for months and all the sudden all these scenes do this. Nothing was changed Even isolating one object this scene does this


I was rendering with these drivers for a long time and all the sudden i cannot even open octaneviewport without an error .

denoiserThread0 : Initialization failed. Restart required
CUDA error 702 on device 0: the launch timed out and was terminated
-> kernel synchronization failed
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to load symbol data to the device(const11)
device 0: failed to upload imager params
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to unregister pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to unregister pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to destroy CUDA event
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to destroy CUDA event
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate device memory
CUDA error 702 on device 0: the launch timed out and was terminated
-> could not get memory info
CUDA error 702 on device 0: the launch timed out and was terminated
-> failed to deallocate pinned memory
CUDA error 702 on device 0: the launch timed out and was terminated

Are you able to PM the scene?

Hi Haze,
it happened even with an empty scene. Now it doesn't. After we got rid of both 2080Ti we tried one at a time. Sometimes it would fail sometimes it wouldn't.
Now I am not sure what to think. Why it didn't work and now it works. is it software or hardware? We had this set up for about a month no problem. The these cuda errors happened and now they went away. Nothing changed.
Can it be 8 pin power cables?

I posted here in details and also attached the video in that post if you want to look at it viewtopic.php?f=80&t=70641&start=10

Thanks

[Goldorak]

Can we have the option for ABC an FBX file to loop, and to have a time offset. This is very handy when using animated ABC files in a large project where sometimes the frame ranges are in high numbers and the looping abc file is only 400 frames long. Hope you guys can consider this.
Cheers

Agreed and already on our internal dev roadmap for 2019.2/3, either through a time envelope node, or some additional pins in the RT animation node.