OctaneRender™ Standalone 3.07

[manalokos]

Thanks Abstrax,

You should consider making the same behaviour for grouped geometry assets.
Usually our workflow to get a geometry asset is to pick the geomtry, select "show in outliner", select the parent in the outliner, and then select the option "show in nodegraph"... it is a waste of time.

[coilbook]

Hi Integrated cloud rendering into each plugin is still a dream that has not been realized for years. Anyway otoy can lift GPU limit to like 25-30 GPUs We want to buy more licenses for other slaves but we've reached our limit of 20. And we do not want to simultaneously render two separate scenes due to too much CPU/mem usage on the master. Thanks

[TRRazor]

Hi Abstrax,
came here to report a bug that had been fixed in the past but must have been re-introduced with one of the last version updates.
When you have a material with a Normal map node, but the normal map image node stays empty, you get a weird darkening-effect with many other strange properties on said surface.

Can this please be fixed? Thanks in advance

[acc24ex]

When are you going to do optimize voxeling so we can fit 100-200% more polygons on the GPUs
- it's obvious it can be done looking at fstorm (I know, you hate it, but they've done just that).. and here you got users wandering what is going on with octane, there was never even a question of whether you can even fit more data - we always assume it's going to get lowered with each new implementation, and here we have a demonstration you can put go 2-3 times more polygons on a gpu ??

Please explain why or how?

[Goldorak]

When are you going to do optimize voxeling so we can fit 100-200% more polygons on the GPUs
- it's obvious it can be done looking at fstorm (I know, you hate it, but they've done just that).. and here you got users wandering what is going on with octane, there was never even a question of whether you can even fit more data - we always assume it's going to get lowered with each new implementation, and here we have a demonstration you can put go 2-3 times more polygons on a gpu ??

Please explain why or how?

Do you remember Octane 1? It was faster and could load more geometry in VRAM than V2 or V3, because it was much, much simpler before.

Brigade engine basically skips Octane's 'voxelizing' for loading and scene changes, so it's nearly as fast as game engine updates. It can load over 2x the geometry buffers than even Octane 1 could The one feature we've shared publicly (so far) about Octane 4 is that Brigade is 100% integrated inside octane, so you will get the best of both engines.

[acc24ex]

When are you going to do optimize voxeling so we can fit 100-200% more polygons on the GPUs
- it's obvious it can be done looking at fstorm (I know, you hate it, but they've done just that).. and here you got users wandering what is going on with octane, there was never even a question of whether you can even fit more data - we always assume it's going to get lowered with each new implementation, and here we have a demonstration you can put go 2-3 times more polygons on a gpu ??

Please explain why or how?

Do you remember Octane 1? It was faster and could load more geometry in VRAM than V2 or V3, because it was much, much simpler before.

Brigade engine basically skips Octane's 'voxelizing' for loading and scene changes, so it's nearly as fast as game engine updates. It can load over 2x the geometry buffers than even Octane 1 could The one feature we've shared publicly (so far) about Octane 4 is that Brigade is 100% integrated inside octane, so you will get the best of both engines.

I remember 0.9alpha
but fstorm can fit 2 times more, not 1 time more actually - so still jumps over brigade .. anyways - where's brigade it's been years

[andw]

Using white balance picker several times (clicking rather frequent) can lead to all GPU failed state

Code: Select all

Started logging on 10.12.17 17:34:27

OctaneRender 3.07 (3070007)

CUDA error 700 on device 0: an illegal memory access was encountered
  -> kernel execution failed(wppick)
device 0: white point picking failed
CUDA error 700 on device 1: an illegal memory access was encountered
  -> failed to load symbol data to the device(imager data)
device 1: failed to upload imager params
CUDA error 700 on device 1: an illegal memory access was encountered
  -> failed to load symbol data to the device(imager data)
device 1: failed to upload imager params

Can I get some more details of this error somewhere? (i.e. hardware/driver fail and so on).

[Kross]

Hi guys, first of all I was thoroughly testing the new version of Octane 3.07, it's great!

However, there are 2 bugs that I discovered in the new version:
1) I have created a new project in Octane 3.07, create a scene with several "Renders Targets", and when I want to save the image from the icon: "Save the current render to disk", "PNG 8-bit", the dialog box to save does not appear... I mean: I can not save the images that I started in Octane 3.07.
This bug does not happen if I open a file created in version 3.06 and try to save from 3.07, ok.
2) If in the space of Nodegraph editor, I select any node "Render Target" and then Ctrl + C and then Ctrl + V to make a copy of the node, effectively makes the copy; but all the modifications do not make them in the copy, but in the original node ... this makes lose all the settings of the original scene. I found a way to avoid this: and it is by selecting any other existing "Render Target" node, and then re-selecting the newly created node. That always happens, and if you do not take care doing this trick, you lose many hours of work.

I await your answers, and thanks in advance!

[abstrax]

Hi Abstrax,
came here to report a bug that had been fixed in the past but must have been re-introduced with one of the last version updates.
When you have a material with a Normal map node, but the normal map image node stays empty, you get a weird darkening-effect with many other strange properties on said surface.

Can this please be fixed? Thanks in advance

This is most likely a problem with the plugin you are using. The normal channel now allows any texture to be used in the normal channel and most likely, the plugin sets it to some RGB colour with value (0, 0, 0) instead of not connecting the pin or setting it to the default normal colour (0.5, 0.5, 1).

If you think that's a general Octane issue, please send me a scene showing the problem. Thank you.

[abstrax]

Using white balance picker several times (clicking rather frequent) can lead to all GPU failed state

Code: Select all

Started logging on 10.12.17 17:34:27

OctaneRender 3.07 (3070007)

CUDA error 700 on device 0: an illegal memory access was encountered
  -> kernel execution failed(wppick)
device 0: white point picking failed
CUDA error 700 on device 1: an illegal memory access was encountered
  -> failed to load symbol data to the device(imager data)
device 1: failed to upload imager params
CUDA error 700 on device 1: an illegal memory access was encountered
  -> failed to load symbol data to the device(imager data)
device 1: failed to upload imager params

Can I get some more details of this error somewhere? (i.e. hardware/driver fail and so on).

I just tried to reproduce it in various scenes, but had no luck.

Does it happen with all scenes? If not, could you send me a scene where this problem occurs? Which GPUs go you use and is it the Standalone you are running?

Thank you.