OctaneRender™ Standalone 2.21.1

A forum where development builds are posted for testing by the community.
Forum rules
NOTE: The software in this forum is not %100 reliable, they are development builds and are meant for testing by experienced octane users. If you are a new octane user, we recommend to use the current stable release from the 'Commercial Product News & Releases' forum.
Post Reply
User avatar
abstrax
OctaneRender Team
Posts: 5506
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

Hi all,

this is the second release of the 2.2 development cycle which - beside lots of bug fixes and smaller features - adds 2 major features:


Light Passes

Light passes let you split the image by the contribution of light sources. Similar to layers, you can assign a light pass ID to emitters and then enable the corresponding pass in the render passes node. There also two additional passes for the sky and the sun. You can find a more detailed explanation here.


Out-of-core Textures

Out-of-core textures allow you to use more textures than would fit in the graphic memory, by keeping them in the host memory. Of course, the data needs to be sent to the GPU while rendering therefore slowing the rendering down. If you don't use out-of-core textures rendering, speed is not affected.

Out-of-core textures come also with another restriction: They must be stored in non-swappable memory which is of course limited. So when you use up all your host memory for out-of-core textures, bad things will happen, since the system can't make room for other processes. Since out-of-core memory is shared between GPUs, you can't turn on/off devices while using out-of-core textures.

You can enabled and configure the out-of-core memory system via the application preferences. For net render slaves you can specify the out-of-core memory options during the installation of the daemon.


Other changes since version 2.20:
  • Restructured the way how post-processing is done when render passes are enabled: If the post-processing pass is enabled, it will not be included in the main beauty pass anymore. You can also specify whether the environment should be included in the post-processing pass or not.
  • Improved machine identification on Windows XP.
  • Multiple UVs work now with subdivision surfaces.
  • Multiple UVs work now with bump/normal mapping.
  • Added new import option to the mesh node, which allows you to create object layer pins directly on the node, without having to use an additional object layer map node.
  • Added multi-connect feature, which is activated when you hold down the CTRL key (CMD key on OSX) while connecting a node with another. If only one connection is possible, this feature is inactive.
  • Added connection "cutter" that kicks in if you hold down the CTRL key and draw a line over the connection to remove.
  • Added pins "F-stop", "Sensor width" and "Focal length", which allow you to control FOV and DOF like on a camera.
  • When a scene node graph (Alembic) is deleted and the corresponding file is not used anywhere else anymore, the file will not be unlocked.
  • Subdivision surface vertex/edge sharpness is now imported correctly from Alembic files (if provided).
  • Fixed crash when loading the animation of a mesh with object IDs and with non-constant topology from an Alembic file.
  • Fixed a couple more crashes during loading of Alembic files, which were reported on the forum.
  • Added a search dialog to quickly find and select nodes and dynamic pins that contain the entered search string. Use CTRL-F to open the dialog.
  • Added a quick search tool to the Lua API browser as well.
  • Added missing render passes and render layer input/output linker nodes.
  • Added option that causes info passes to be tonemapped or not (in the past they were always tonemapped). That's currently more a limitation since for info passes we have to make that decision during render time unlike the beauty passes, which are tonemapped at a later stage.
  • Added option "visibility only" to the render layer node, which will switch the behaviour to affect the primary ray visibility. As a consequence, the shadow and reflection layer passes don't work in that mode.
  • Added stereo panoramic rendering. This is experimental and we will write something about this in the near future.
  • Brought the option back to the pass save dialog, which allows you save passes as tonemapped .exr files.
  • When dithering is enabled in the camera imager and the rendering is saved as untonemapped HDR image, dithering is applied to this result, too. Be aware though that it is not a good idea to change the brightness of an HDR image with dithering after being saved, since this will also stretch the dither noise.
  • If a material map or object layer map node is connected to a geometry node other than a mesh node, all dynamic input pins are removed.
  • Value changes via the scroll wheel are now merged to 1 undo step.
  • Fixed incorrect undo when a node is created via the node graph editor.
  • Fixed missing support for .exr files in alpha image texture nodes.
  • Fixed polygon side texture that was broken, when a bump or normal map is used (see http://render.otoy.com/forum/viewtopic. ... 18#p216818).
  • Fixed numeric fields of the color picker on Mac OSX.
  • Fixed emission in scattering media (see http://render.otoy.com/forum/viewtopic. ... 39#p217839).
  • Fixed render reset when a render region is applied to an animated scene.
  • Fixed camera reset command for panoramic camera nodes.
  • Fixed incorrect update of render passes while rendering is paused.
  • Fixed scan of Lua script headers that don't have the header starting directly in line 1.
  • Fixed frame numbering in the render animation script when only a subset of the frames is rendered (see http://render.otoy.com/forum/viewtopic. ... 80#p218780).
  • If a Lua script is run from the command line and it returns a numeric value, this value will be set as the application exit status. If any other error occurs the returned status code is 100.
  • Scripts that are stored in a Lua script graph are not automatically converted into a file anymore.
  • Fixed crash when a scripted node graph tries to access the scene root graph, while loading the project.
  • Running a script via the Lua editor disables the UI similar to the case when it's run via the script menu.
  • Added octane.node.getAnimator(), octane.node.getAnimatorIx(), octane.nodegraph.getAnimator() and octane.nodegraph.getAnimatorIx() to the Lua API to allow Lua scripts fetch animated attributes.
  • Added octane.render.getRenderRegion() and octane.render.setRenderRegion() to control the current render region.
  • Creating a node via a Lua script will give the node its default name if no name is specified.
  • Fixed the problem in network rendering that saving an HDR image can block the slave too long and the daemon kills the process (http://render.otoy.com/forum/viewtopic. ... 22#p218822).
  • Fixed a crash in the net render slave that occurred sometimes when render passes are used.
  • Fixed a crash that occurred sometimes when you stop the daemon via CTRL-C while the slave is running.

Quickfix release 2.21.1

We updated the release builds to version 2.21.1 which fixed these 3 annoying issues:
  • Fixed vertically mirrored UV maps when subdivision surfaces is used.
  • Fixed broken material/texture preview.
  • When aperture is set to 0 it will be kept at zero, when other camera parameters like the focal length are changed.

Download

Windows
- 64-bit installer
- 64-bit ZIP archive

Mac OS X
- 64-bit DMG

Linux
- 64-bit ZIP

Happy rendering,
Your OTOY NZ Team
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
prehabitat
Licensed Customer
Posts: 495
Joined: Fri Aug 16, 2013 10:30 am
Location: Victoria, Australia

WoaH!
Win10/3770/16gb/K600(display)/GTX780(Octane)/GTX590/372.70
Octane 3.x: GH Lands VARQ Rhino5 -Rhino.io- C4D R16 / Revit17
User avatar
smicha
Licensed Customer
Posts: 3151
Joined: Wed Sep 21, 2011 4:13 pm
Location: Warsaw, Poland

Morning coffee and 2.21 :) Thank you!
3090, Titan, Quadro, Xeon Scalable Supermicro, 768GB RAM; Sketchup Pro, Classical Architecture.
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540
Tugpsx
Licensed Customer
Posts: 1150
Joined: Thu Feb 04, 2010 8:04 pm
Location: Chicago, IL
Contact:

Nice, more new toys to play with
Win 11 64GB | NVIDIA RTX3060 12GB
User avatar
smicha
Licensed Customer
Posts: 3151
Joined: Wed Sep 21, 2011 4:13 pm
Location: Warsaw, Poland

Marcus,

Could you please say something more about out of core textures RAM allocation.

1. When will I see the system RAM used? I just added the 200MB texture to a scene with the 'out of core option' on and 'out of core used' says 0.

2. GPU head room - what exactly it does? For a GTX Titan is it better to set it to 1024?

3. How can I control of how much of vram is taken by textures?
Attachments
mem.jpg
3090, Titan, Quadro, Xeon Scalable Supermicro, 768GB RAM; Sketchup Pro, Classical Architecture.
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540
User avatar
abstrax
OctaneRender Team
Posts: 5506
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

smicha wrote:Marcus,

Could you please say something more about out of core textures RAM allocation.

1. When will I see the system RAM used? I just added the 200MB texture to a scene with the 'out of core option' on and 'out of core used' says 0.
When (availableGpuRam - oocHeadroom) is not enough to facilitate a texture on the GPU, it will be stored in host RAM insted. In other words: Only when a device runs out of memory you will see out-of-core memory being used.
2. GPU head room - what exactly it does? For a GTX Titan is it better to set it to 1024?
That's the amount of space you reserve for other GPU users.
3. How can I control of how much of vram is taken by textures?
Octane maximizes the amount of memory up to the headroom limit you specify. In general, you want to have the maximum amount of stuff on the GPU since this memory is fast.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
User avatar
smicha
Licensed Customer
Posts: 3151
Joined: Wed Sep 21, 2011 4:13 pm
Location: Warsaw, Poland

Another question: if I have e.g., 5GB of obj files and 4GB of textures and let's say a titan can allocate up to 6GB - will octane load 5GB of objs and 1GB of textures to vram and excessive 3GB of textures to RAM?
3090, Titan, Quadro, Xeon Scalable Supermicro, 768GB RAM; Sketchup Pro, Classical Architecture.
Custom alloy powder coated laser cut cases, Autodesk metal-sheet 3D modelling.
build-log http://render.otoy.com/forum/viewtopic.php?f=9&t=42540
User avatar
abstrax
OctaneRender Team
Posts: 5506
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

smicha wrote:Another question: if I have e.g., 5GB of obj files and 4GB of textures and let's say a titan can allocate up to 6GB - will octane load 5GB of objs and 1GB of textures to vram and excessive 3GB of textures to RAM?
In principle yes, but the maximum amount of geometry memory you can use is currently limited to ~19 million triangles, which will use about 3GB of memory and a 5GB OBJ file will use a lot less memory than 5GB on the GPU.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
User avatar
aLeXXtoR
Licensed Customer
Posts: 294
Joined: Sat Oct 29, 2011 5:38 pm
Location: Russia, St. Ptetersburg
Contact:

Amazing!
Win 10 Pro 64bit | GeForce RTX 2070 SUPER 8Gb | AMD Ryzen 7 3700X | 32GB RAM
WebSite
Showreel
ratata
Licensed Customer
Posts: 74
Joined: Tue Aug 19, 2014 1:28 am
Location: Ljubljana, Slovenia
Contact:

Awesome news!

Will there be any official info on the speed hit that off core memory will be? Examples like 1GB, 2GB...

It would be interesting to know if now the most cost / value GPU would be 780 ti 3GB, in cases that you go over 3GB limit only here and there...
http://www.twitter.com/sandidolsak

6850k|64gb|4 x GTX980ti
Post Reply

Return to “Development Build Releases”