OctaneRender™ Standalone 3.00 alpha 3

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
stratified
OctaneRender Team
Posts: 945
Joined: Wed Aug 15, 2012 6:32 am
Location: Auckland, New Zealand

grimm wrote:The environment medium doesn't work for me. :( I made this test scene in Blender to test using the sphere around the camera method. Here is what I was able to get with scattering set to 0.0002 and scale set to 80:
The attachment fog-testv3.png is no longer available
Here are the settings in the Blender plugin:
The attachment fog-test-blender.png is no longer available
Then when I tried to replicate the same settings in V3 standalone I get this:
The attachment fog-testv3-octane.png is no longer available
Here are the files:
The attachment fog-test2.blend is no longer available
The attachment v3-fog-test.ocs is no longer available
Thanks :)
I had a look at this, the radius of your medium is too small (it's set to 3 in your project). It should be something very large. I think you used the radius of the sphere around the camera?

This is how it works, the enviroment medium is defined as a sphere around the camera:
How the environment medium works.
How the environment medium works.
cheers,
Thomas
User avatar
grimm
Licensed Customer
Posts: 1332
Joined: Wed Jan 27, 2010 8:11 pm
Location: Spokane, Washington, USA

Thanks for looking into that Thomas. That was the issue, I thought it worked similar to the inverted normals sphere around camera method. :) You are right about setting it really large, for this scene I had to set the sphere size to at least 600. Any lower and very strange artifacts would start to be rendered.

Jason
Linux Mint 21.3 x64 | Nvidia GTX 980 4GB (displays) RTX 2070 8GB| Intel I7 5820K 3.8 Ghz | 32Gb Memory | Nvidia Driver 535.171
User avatar
abstrax
OctaneRender Team
Posts: 5509
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

Notiusweb wrote:Here is a video of the multi-GPU viewport lag I had referenced. It compares V3 vs V2.

https://vimeo.com/152060121
The next version will have a workaround for an issue that occured when Octane tried to page-lock host memory for usage on the GPU. For some reason this operation got blocked by kernel calls on other devices. This issue affects mainly tonemapping and I think it's what you experience on your system, since you have many devices. Let's hope that the workaround fixes your problem.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
User avatar
Notiusweb
Licensed Customer
Posts: 1285
Joined: Mon Nov 10, 2014 4:51 am

The next version will have a workaround for an issue that occured when Octane tried to page-lock host memory for usage on the GPU. For some reason this operation got blocked by kernel calls on other devices. This issue affects mainly tonemapping and I think it's what you experience on your system, since you have many devices. Let's hope that the workaround fixes your problem.
Great, thank you Sir!
Win 10 Pro 64, Xeon E5-2687W v2 (8x 3.40GHz), G.Skill 64 GB DDR3-2400, ASRock X79 Extreme 11
Mobo: 1 Titan RTX, 1 Titan Xp
External: 6 Titan X Pascal, 2 GTX Titan X
Plugs: Enterprise
User avatar
Notiusweb
Licensed Customer
Posts: 1285
Joined: Mon Nov 10, 2014 4:51 am

Hello all, this is a summary list I made of the posts from Alpha 1, 2, 3, broken into categories of Issues, Wishlist, Developer Notes, and then my own notes for the posts. I put the user name at the end in parenthesis.
It might help see what was going on with user encountered issues (Issues), see who else had requested something (Wishlist), or get up to speed on the V3 development within context of release (Developer Notes). Also, if you see your name by an issue, and its not an issue or you have a workaround, you could note it...
I had fun reading this all, hopefully I didn't misquote anyone! Wishlist and Developer Notes are more fun to look at :ugeek:

OCTANE RENDER 3 STANDALONE ISSUES

Alpha 1
1 - Just like any previous version it's 4-8 times slower on OSX. Slower rendering of character animation, cloth sims, point cached animation on OSX. ( milanm)
2 - every time I try to activate my account on the v3 standalone it says it doesn't recognise the details, although when I login using the same details on the Otoy site it works fine. ( MikeJCTyler)
3 - I made a couple of manipulations scattering in the sun, function does not work Archived packages, also I saw these bands with fog, if you move the slider to the scale of octane Freeze for 2-3 minutes, video driver stopped responding after displacement, the scale is broken (using NVIDIA 355.82) (Elvissuperstar007 )
4 - getting 20-40 % slowdown in OR3 compared to OR2.24 on simple demo scene, Insane memory consumption compared to 2.42 for the same scenes. ( uncia)
5 - Octane 2 uses up 10MB for some simple cubes, while the same scene in Octane 3 uses 350MB. If I set parallel samples to 16, it uses 690MB. This is a massive difference (35 times greater mem usage than octane 2 using defaults) (funk)
6 - I noticed that many times rendering stops updating, but updates imediately when pausing. Also it doesn't updates sometimes when changing parameters. ( manalokos)
7 - It looks like IES lights are broken. Simple cube and a single IES light. Version 2.4 renders fine, version 3 shows almost no light (wimvdb)
8 - Volumes cause light to enter inside closed spaces. I attached an image illustrating what happens inside a closed box when a volume intercepts the surface. (manalokos)
9 - Using the built in environment medium node with a color other than white (ie. blue like atmospheric haze), I get a lot of strange hotpixels (that the hotpixel setting won't fix). ( profbetis)
10 - Not sure if this is a known issue, but info passes are being tonemapped even when the tonemap toggle is off. I have my imager gamma set to 2.2, with info pass tonemap off.
- In octane 2 the geometric normal pass has deep blues and background is black.
- In octane 3 the geo normal pass is tonemapped and shows light purples with grey background (funk)
11 - I also noticed this change if info pass distributed ray tracing is enabled
- In octane 2, if I enable the motion vector info pass, it disabled motion blur.
- In octane 3, the motion vector pass actually gets motion blurred (this is completely wrong) (funk)
12 - when OOC feature is used (a bit slower than the 2.x), As the kernel has changed, I have a system crash when I use with all my 8 GPUs. (itou31)
13 - The volume loader drops all of it's settings if you save and open your octane file. (reverts to default)
14 - If you have loaded some animated alembic files in your scene, and add volume sequence using the animation option in the loader, it changes the whole timeline to that. That does not seem right. (gardeler)
15 - as you bump resolution up & up You see multiple GPUs utilised worse & worse. Let's say if You rendering 1kx1k output.. three local GPUs are 99% load 99% percent of time.. however..if You push resolution higher.. one GPU stays fully loaded all the time, while others keep dropping down.. & You see those 0% GPU loads for longer & longer periods (Glimpse)..It's the tone mapping which blocks rendering, even if only GPU does the actual tonemapping. That's why you can now increase the tonemap interval to up to 120s to reduce the overhead. I haven't looked into this yet in detail so there may be a different issue, too. (Abstrax)
16 - with 12 GPU that the view-port display has choppiness . as I reduce the # of GPUs towards 1, the display linearly smooths up. effect occurs despite the fact that I have a dedicated card for display not used in the rendering. linear increase in display choppiness does not occur in 2X build, in any range between 1 and 12 active GPUs, it is always smooth. (Notiusweb) The next version will have a workaround for an issue that occurred when Octane tried to page-lock host memory for usage on the GPU. For some reason this operation got blocked by kernel calls on other devices. This issue affects mainly tonemapping and I think it's what you experience on your system, since you have many devices. (Abstrax /Alpha 3 post)

Alpha 2
1 - fix the spikes on some meshes when subdividing the object in version 2 and 3 (Zay)
2 - differences in the glossy reflections (RobSteady) I think the differences are due to the slightly different behaviour of bump mapping. If you remove the bump maps, it should render exactly the same. Otherwise there is a bug somewhere. (Abstrax)
3 - script for simple animation paths...With 3.00 it doesn't work anymore (cannot set output folder, and render button doesn't work) (ff7darkcloud)
4 - meshes with displacement applied render out a black texture. (Rikk The Gaijin)
5 - the AO render pass has limit of 1024, which is still very noisy. (Rikk The Gaijin)
6 - I imported a mesh with 43 million polys, and Octane closed down without warning. (Rikk The Gaijin)
7 - tangent normal render pass output a flat RGB 128,128,256 color image. (Rikk The Gaijin)
8 - Object layer works only if the render target node is connected (Smicha)
9 - Looks like displacement is killing the motionblur (gardeler). That is a known issue, we have not found a way to make triangles with a displacement render correctly with motion blur without artifacts (Roeland)
10 - Octane 3 Alpha 2 crashes when saving an image while using network rendering, when slaves are still finishing their samples. (manalokos)
11 - testing SSS and plugged an emission texture into it with "cast illumination" disabled. It's still casting light on the object around it though. I tried this in octane 2 and see the same issue. (funk)

Alpha 3
1 - when 'O' for picking mode is used to selected any geometry that has an active object layer pin, all settings for obj are gone (only the object layer is shown) - see the screenshot. I have to manually select an obj node from the node graph to access geometry settings. (smicha)


OCTANE 3 WISHLIST

Alpha 1
1 - Would it be possible to implement in version 3 a feature which would compile only objects that are being added to the scene ? For instance if I compiled a scene with 10 million polygons and I need to add a 10 poly object currently I need to reload entire scene which takes quite a while. ( BorisGoreta) ( Seekerfinder) (FrankPooleFloating) (smicha)

Alpha 2
1 - i'd like user adjustable priority to get back in octane very much. (Lewis)
2 - hotkey which could toggle an 'active GPU profile' without having to pause the render, go to File>Preferences>Devices, and then select GPUs by clicking the little check boxes. (Notiusweb)
3 - If parallel samples / max tiles, could be adjusted without restarting the render, then we wouldnt need render priority (funk).
4 - if Dev Team can make rendered images more crystal clear. or let us give more settings to make that happen (oguzbir)
5 - atmospheric effects like bluish objects that are far away, can it be added to the Octane Sun where we can control fog, haze etc. (coilbook)
6 - possible to alter or add the functionality of Random Color Texture to the elements of the objects.(oguzbir)

Alpha 3
1 - V3 manual and downloadable version (prehabitat)
2 - export to JPEG option (Notiusweb)
3 - ability to save renders with any loaded background image (Notiusweb)
4 - sub. div. level pin. , e.g., in the object layer pin or anything else on the obj node. Or at least a global button in a toolbar to turn on/off all sub divs. Same for displacement - global on/off button would be so helpful. (smicha)
5 - render priority (Techie) (FrankPooleFloating) We will have a look at how we can implement render priority in the new system. Don't know yet how it can be done. (Abstrax)
6 - seems the volume step length in the kernel setting makes a huge impact on how an imported VDB actually looks like. It seems currently the volume step length can only be set on a scene level. Is there any chance that the volume step length can be set on an object level in the future? (livanchene)
7 - support for OpenCL (Yan) (smicha)


DEVELOPER NOTES

Alpha 1
1 - In simple scenes where you've got plenty of VRAM left: Increase "parallel samples" to the maximum.
In complex scenes where VRAM is spare: Set it to the highest value without running out of memory. It should usually still be faster than before or at least render with roughly the same speed. (Abstrax)
2 - It's not yet possible to add noise to volumes, or to convert a mesh to a volume inside Octane. We might introduce these later in 3.x since OpenVDB's sdk has some special features for these, but, as for now, we're on a tight schedule to get 3.00 stable. (Haze)
3 - V2+V3 Upgrade will be EQUAL to outright purchase of V3. (confirmed by Goldorak)
4 - Baking is a well known lighting tool in games, and also used in in archvis walkthroughs. If you have a static scene, it allows you to store all the GI in textures and move around in real time, even with very simple game engines.
The more sophisticated work we are doing with the ORBX player down the line will take this to another level through light field and volume rendering. This will allow you to move with 6DOF, with reflections, lighting/object layers and support for almost anything that Octane 3 can render (volumes, hair for example). (Goldorak)
5 - OSL support may give you greater procedural control over BRDF material properties, and possibly allow you to create completely new BRDFs (we will support this only if we can validate the shader is energy conserving). In Brigade, where this code was first implemented, we can confirm much of this is already working - for example GTX BRDF materials can be added to Brigade via pure OSL. So we'll see how much of this can be brought into Octane 3 (which has different constraints than Brigade) once OSL is further along in the V3 beta. (Goldorak)

Alpha 2
1 - Jim Star is currently developing an important part of 3.x that we are planning to introduce in Max/Maya/Blender/SU first before rolling it out to all plug-ins. (Goldorak)
2 - can I pre-purchase a license for c4d plugin(3.x) ? cuz I need to buy a new c4d plugin license(right now it's still 2.x and 199euros) but I don't want to pay an extra upgrading price(say 99 euros?) in the future??? I need 3.x license(assume 3.x will be 199 euros too?) (c4d) No, the v3 final cost will roughly be like v2 + v3 upgrade, so there is no difference in buying it now or later on in terms of costs, but if you buy it now you can already use the new v3 features during the development (bepeg4d)
3 - beside playing with the medium node, (try) to reduce or enlarge, based on your scene, the Medium radius value in the Environment node...It has a great impact on the deepness of the effect (bepeg4d)

Alpha 3
1 - in the long run it will be possible for 3rd parties to create plugins for OctaneRender standalone (confirmed by stratified)
2 - OpenSubdiv 3, set in mesh node (preferences, the "wrench" icon) (Abstrax)
3 - OpenCL has been on hiatus for quite a while and we will continue working on it after OSL has been done, i.e. it may become available around middle of the year. Please be aware, that there is a chance that we may run into insurmountable problems and we can't get it working (Octane is not a simple path tracer anymore). I didn't run into any issues during my initial tests, but everybody tells me how bad it is, so we will have to see. There is also a small possibility that AMD's CUDA implementation as part of AMD's Boltzman initiative could be used instead of an OpenCL implementation, but we will see, when we get there. Most likely it will not be useful for Octane. (Abstrax)
4 - We are still busy working on major feature enhancement to the volumetric support you're seeing in alpha 3. The volume node will look substantially different very soon, but in terms of priority, colour ramps for volumes will be done before the step length enhancement. Some more information on the step length: We use raymarching for rendering volumes, and each step along the march is a fixed size in metres. This is the step length you are seeing in the kernels. The smaller the step length, the more accurate the results. When you have a larger step length, one that goes across many voxels, Octane must pick a single density for the whole step. Many voxels will get missed, and when it misses a few voxels with a high temperature, then the emission will be lower in that area. This is why lowering the step length changes the look of a volume. The same goes for absorption and scattering. (Haze)
5 - Working offline is not possible with v3 alpha beta test versions. You have to use v2 if you need to work offline (bepeg4d)

USER NOTES

1 - Render Speed difference between V2 and V3 appears to be understood as a trade-off:
- Lower res/ram scenes slower compared to V2, much higher res/ram scenes faster compared to V2
- V2 faster with Direct Lighting vs V3, V3 faster at PT and PMC vs V2 (Notiusweb)
2 - Many Volume / VDB discussion / tutorials are in Alpha 1 (Notiusweb)
Win 10 Pro 64, Xeon E5-2687W v2 (8x 3.40GHz), G.Skill 64 GB DDR3-2400, ASRock X79 Extreme 11
Mobo: 1 Titan RTX, 1 Titan Xp
External: 6 Titan X Pascal, 2 GTX Titan X
Plugs: Enterprise
User avatar
Seekerfinder
Licensed Customer
Posts: 1600
Joined: Tue Jan 04, 2011 11:34 am

Wow Notius. Thanks (I think)! You must have Tutor as your "thoroughness"... tutor.

Of all the issues listed it's the following observation that worries me most.
Notiusweb wrote: 4 - getting 20-40 % slowdown in OR3 compared to OR2.24 on simple demo scene
I have not tested this and I don't know if it's verified. But if true it would be at least the 3rd time that we're losing significant speed (including the fermi-kepler transition) since Octane 1 (we got some speed restored with the PT kernel in 2 which is great). The focus seems to be all on features. Great. We need them. But it's like original focus of what Octane was all about is being lost as a priority. By all accounts GPU programming seems to be real hard and a lot of what's being developed sounds great (looking forward to PS & UE4 integration). But all those 15-25% slowdowns add up. Does feature partity with a CPU renderer like Maxwell eventually lead to speed parity with those renderers as well?

I know there's no simple answer here and I still love Octane. But I'd love to see an Octane (not Otoy) mission statement...

Best,
Seeker
Win 8(64) | P9X79-E WS | i7-3930K | 32GB | GTX Titan & GTX 780Ti | SketchUP | Revit | Beta tester for Revit & Sketchup plugins for Octane
User avatar
abstrax
OctaneRender Team
Posts: 5509
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

Seekerfinder wrote:Wow Notius. Thanks (I think)! You must have Tutor as your "thoroughness"... tutor.

Of all the issues listed it's the following observation that worries me most.
Notiusweb wrote: 4 - getting 20-40 % slowdown in OR3 compared to OR2.24 on simple demo scene
I have not tested this and I don't know if it's verified. But if true it would be at least the 3rd time that we're losing significant speed (including the fermi-kepler transition) since Octane 1 (we got some speed restored with the PT kernel in 2 which is great). The focus seems to be all on features. Great. We need them. But it's like original focus of what Octane was all about is being lost as a priority. By all accounts GPU programming seems to be real hard and a lot of what's being developed sounds great (looking forward to PS & UE4 integration). But all those 15-25% slowdowns add up. Does feature partity with a CPU renderer like Maxwell eventually lead to speed parity with those renderers as well?

I know there's no simple answer here and I still love Octane. But I'd love to see an Octane (not Otoy) mission statement...

Best,
Seeker
The mission statement is the same is has always been: Our priorities are stability, performance and features. In that order. Please do your own tests first and don't jump on stuff like "Octane is getting 20-40 % slowdown" without testing yourself.

That's not true until proven. As I have written and discussed multiple times already, in simple scenes you have to bump up the parallel sample count and max samples per tile settings. Yes it requires more VRAM, but simple scenes don't require a lot of VRAM themselves so it doesn't matter. Anything non-trivial is most of the times faster than v2 (not always, but that's just the nature of software development). Sometimes even up to 1.5-2x faster. Most of the production scenes that have arrived in support were far from simple and usually improve in render speed compared to v2.

And no, we haven't continuously lost performance. During the v2 cycle we actually gained performance in many scenes due to the introduction of better sampling and the coherent mode/ratio. So please check your facts first, before you make statements like these.

EDIT: And in v2 we added additional options to clamp fireflies (GI clamp) and to soften caustics (caustic blur) which both help you to achieve cleaner renders faster...
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
User avatar
stratified
OctaneRender Team
Posts: 945
Joined: Wed Aug 15, 2012 6:32 am
Location: Auckland, New Zealand

grimm wrote:I'm not sure that this has worked for a while but the obj file loader doesn't read the mtl file correctly. It used to just have problems with normal maps but now it doesn't read any texture type correctly.
fog-test2-nofog.obj
The forum software doesn't let me upload the mtl file but here it is:

Code: Select all

# Blender MTL File: 'fog-test2-nofog.blend'
# Material Count: 2

newmtl Material.001
Ns 96.078431
Ka 1.000000 1.000000 1.000000
Kd 0.640000 0.640000 0.640000
Ks 0.500000 0.500000 0.500000
Ke 0.000000 0.000000 0.000000
Ni 1.000000
d 1.000000
illum 2

newmtl Material.002
Ns 96.078431
Ka 1.000000 1.000000 1.000000
Kd 0.640000 0.640000 0.640000
Ks 0.500000 0.500000 0.500000
Ke 0.000000 0.000000 0.000000
Ni 1.000000
d 1.000000
illum 2
These were generated by Blender.
Unless I'm missing something, loading this MTL file works as expected...

There are no image textures described in this MTL file (via directives map_ka, map_kd, map_ks, map_ns,, map_bump, bump, map_d). The 2 materials are identical and Octane correctly creates a glossy material for them using:
  • Kd 0.640000 0.640000 0.640000 -> diffuse RGB texture with value (0.64, 0.64, 0.64).
  • Ks 0.500000 0.500000 0.500000 -> specular RGB texture with value (0.5, 0.5, 0.5).
cheers,
Thomas
User avatar
abstrax
OctaneRender Team
Posts: 5509
Joined: Tue May 18, 2010 11:01 am
Location: Auckland, New Zealand

cfrank78 wrote:...
By the way - the offline activation block for alpha and beta is one of the worst YOU HAVE TO BUY IT, GIVE US OUR MONEY ACTION ive seen in a long, long time!

But hey - just one little point more on the list why V3 is a dissappointment for some of us.
Why? You can use the v3 during the alpha/beta period without paying for the upgrade. Nobody asks for money here. And no you don't have to buy the upgrade afterwards. That you don't get the upgrade for free should be obvious. Sometimes I really wonder what you guys are smoking...
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

abstrax wrote:Sometimes I really wonder what you guys are smoking...
8-)
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
Post Reply

Return to “Development Build Releases”