Page 4 of 10

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 1:15 pm
by SergKlyosov
I also tested my scene, which takes Octane to collect all nodes and start rendering - 5.5 min
In "Another render" it took 18 sec to collect everything and start actual render process


Paride, NeonZorglub

As you can see it is not GPU VRAM speed limitation, the issue is specific to Octane. So developers need to to optimization of loading algorithm

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 2:26 pm
by paride4331
RobSteady wrote:
paride4331 wrote: > 15.800.000 polys > GPU mem 3.85 GB 4554 objs time to voxel > 20 secs.
When Octane V4 will be released, I think there won't be much more for any of render engines.
Meanwhile the competition stays lightyears ahead. This is “another gpu renderer“ on the market:
04b.png
Hi RobSteady,
I know, the grass is always greener on the other side of the fence.
Said that, this is an experimental version in development. Could you test your scene using SA V4 XB4? what about?
Regards
Paride

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 3:08 pm
by clicky
Trying to set some scene in V4, because in V3 its not possible because of the out of core limit.
Its a pretty large scene with about 25 UDIMs and displacements and such

After a while I ran into this error:
Render failure in Viewport Rendering: GPU Failure
this is from the log:
Tried to access attribute via invalid ID A_OUT_OF_CORE_TEX_VRAM_LIMIT
Tried to access attribute via invalid ID A_OUT_OF_CORE_GEO_VRAM_LIMIT

Seems like there still is some out of core limit, or not?

screenshot in attachment, anyone know what this may be?

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 3:19 pm
by RobSteady
Hi Paride,

unfortunately not. I left Octane when v3 came out and materials from v2 did not work anymore (or was it 2.something, don't remember). We had already converted hundreds of assets and projects that we need to revisit. By that time there was no autoamtic conversion available for a long time.
Besides that lots of empty promises...

So, good luck everyone!

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 3:28 pm
by ramone163
paride4331 wrote:
RobSteady wrote:
paride4331 wrote: > 15.800.000 polys > GPU mem 3.85 GB 4554 objs time to voxel > 20 secs.
When Octane V4 will be released, I think there won't be much more for any of render engines.
Meanwhile the competition stays lightyears ahead. This is “another gpu renderer“ on the market:
04b.png
Hi RobSteady,
I know, the grass is always greener on the other side of the fence.
Said that, this is an experimental version in development. Could you test your scene using SA V4 XB4? what about?
Regards
Paride
Hi Paride.
Have you tried to render this scene from Fstorm with 68 mil poligons?
Octane 4 and Octane4+ 3ds max just crash with my 1080+1080Ti
https://www.dropbox.com/s/7fosk3sv6lylg ... s.zip?dl=0

If we are talking about this numbers:
Octane 26 mil polygons = 6,3 gb and 6 min to process
Fstorm 69 mil polygons = 4,4 gb and 2 min to process
Octane need not just optimization, it needs HUGE optimization.

PS. About "grass is always greener". This is unfair. You can always say that this is experimental build or this is the final release, but all will be fixed in version 4.1 and so on.

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 3:38 pm
by paride4331
HHbomb wrote:Hi,
in XB4 6.04, is "Min. denoiser samples" work ? I can change value, octane begin denoising at different values.

- if I set 50, octane begin to denoise when I see Smp/px 1024/1024/16000 ( because render is very fast )
- if I set 50, and then set Max denoiser interval to 1s it begin to denoise at 112 Smp/px, but then octane denoise at each 1s...

Need a fix ?

Hi HHbomb,
unchecking "denoise on completition":
first denoise start on "min denois sampl" samples.
following denoising start on "Max. denois interv" (in seconds) when "min denois sampl" is false.
E.g.
"min denois sampl" > 500 (sampl)
"Max. denois interv" > 3 (sec)

start render >
first denoise starts on 500 samples
second denoise starts after 3 seconds if render samples is not 1000
and so on.
Regards
Paride

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 4:04 pm
by ramone163
Hi Paride.
Problem with XB4.
Model without TurboSmooth. Everything is fine. fast moving in the viewport and in render view.
Same model with TurboSmooth, fast moving in the viewport. But very slow moving when rendering. And render have some artifacts, cuts the image.
Check my video.

PS. I think I found where the problem is. Slow viewport and render artifacts only in Isolate Mode. In normal mode everything is ok.

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 5:16 pm
by HHbomb
Hi, could we have ability to add an Object layer helper by other way than opening the octane object property ? Like a light, a volume or by script ?
When an object is excluded from the sun, it is no more lighted by the sun, but it always generate shadows from the sun ? Is it normal behavior or not ?

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 6:42 pm
by coilbook
Hopefully octane 4 will fix this slow processing time. For example, this is vray with a very large VDB Processing time 2 seconds https://youtu.be/-_ZCWOq9T1o?t=1m34s

Re: OctaneRender® for 3ds max® v4.00-XB4 - 6.04 [TEST]

Posted: Thu Aug 09, 2018 6:45 pm
by coilbook
Meanwhile the competition stays lightyears ahead. This is “another gpu renderer“ on the market:
04b.png
[/quote]

Hi RobSteady,
I know, the grass is always greener on the other side of the fence.
Said that, this is an experimental version in development. Could you test your scene using SA V4 XB4? what about?
Regards
Paride[/quote]

Hi Paride.
Have you tried to render this scene from Fstorm with 68 mil poligons?
Octane 4 and Octane4+ 3ds max just crash with my 1080+1080Ti
https://www.dropbox.com/s/7fosk3sv6lylg ... s.zip?dl=0

If we are talking about this numbers:
Octane 26 mil polygons = 6,3 gb and 6 min to process
Fstorm 69 mil polygons = 4,4 gb and 2 min to process
Octane need not just optimization, it needs HUGE optimization.

PS. About "grass is always greener". This is unfair. You can always say that this is experimental build or this is the final release, but all will be fixed in version 4.1 and so on.[/quote]

I agree. What's the point of a super fast multi GPU rendering or AI rendering in seconds when processing takes 3-4 min per frame. And scenes that have multiple phoenix grids take even longer . This should be number one priority for otoy.