Page 1 of 6 123 ... LastLast
Results 1 to 15 of 81

Thread: GPU Rendering - Anywhere in the development pipeline for Lightwave?

  1. #1
    Registered User
    Join Date
    May 2012
    Location
    Virginia
    Posts
    428

    GPU Rendering - Anywhere in the development pipeline for Lightwave?

    I promise, I am not trying to beat a dead horse with this thread.

    I'm getting ready to make some significant team software purchase decisions; specifically, to purchase Octane Render for a rather large team (because bless Lightwave, but the render times are killing us).

    There are many reasons why I would prefer not to do this that I'm sure many of you are very familiar with; fiscal reasons, learning curve reasons, etc.

    I know that this subject has been brought up many times and that requests for development roadmaps are largely ignored by the development group; but it would be irresponsible of me to commit this amount of money on software without at least trying to gather as much information as possible.

    Especially in this case, where the purchase would take the place of other software or hardware.

    In addition, I feel like asking this question is even more important now that Lightwave has shifted to a much more aggressive development cycle.

    So here goes.

    1. Is there any intent on the part of the development team to implement GPU rendering in Lightwave?


    1. If not, is there any intent to expand plug-in compatibility / support to any other third-party GPU renderers (Redshift, Vray, Iray, Thea, etc.)?


    We are sticking with Lightwave, regardless; but if there is the possibility that we don't have to commit to this purchase right now, I'd just like to know.

    Thank you for any information you might be able to provide and I apologize in advance for bringing up this subject again.

  2. #2
    Oh boy, I know that feeling.
    I was reading Octane's updated roadmap where it is ALL subscription. Eww... Not really interested in that.

    You won't get an answer here on the forums, I would bet. But it is essentially the question in my mind today after the above read.

    Joining in for the support of such a notion in LW.
    Robert
    Robert Wilson, MA Deaf Ed.
    Indiana Deaf School
    This e-mail may contain information protected under the Family Educational Rights and Privacy Act.

  3. #3
    Registered User
    Join Date
    Mar 2016
    Location
    Oxford, UK
    Posts
    794
    For business critical decision purposes you have to assume the answer is No, because what might happen can't be allowed to influence that. In practise I think all sorts of things go on behind the scenes, some of which come to fruition and some which don't (which probably explains why it all goes on behind the scenes ).

    This question came up around a week ago at the Lightwave Round Table discussion and a NT dev said that they were in talks about AMD Prorender, but no time frame.....but please don't take that to mean that AMD Prorender has somehow been promised after a wait. Dev made point that LW is used in so many scenarios they consider all of those, but NT are aware of the desire to see GPU in LW. I don't know what the practicalities would be of bolting Blender to your pipeline and consider Cycles till something more concrete happens.

    Good luck anyway.

  4. #4
    Yeah you can't make a financial or time investment in a hypothetical roadmap even if they were to provide you one.
    And personally I wouldn't base one off Lightwave development which has changed direction like the wind blows.

    That said, all of my investment on the rendering side of things is geared towards GPU.

  5. #5
    Registered User
    Join Date
    May 2012
    Location
    Virginia
    Posts
    428
    Quote Originally Posted by TheLexx View Post
    For business critical decision purposes you have to assume the answer is No, because what might happen can't be allowed to influence that. In practise I think all sorts of things go on behind the scenes, some of which come to fruition and some which don't (which probably explains why it all goes on behind the scenes ).
    True. I only asked because I have some time before the situation reaches critical mass. We are ramping up some aspects of our 3d development that we hadn't previously had a need for before, which is why I needed to start the process to make sure we can deliver.

    Quote Originally Posted by TheLexx View Post
    This question came up around a week ago at the Lightwave Round Table discussion and a NT dev said that they were in talks about AMD Prorender, but no time frame.....but please don't take that to mean that AMD Prorender has somehow been promised after a wait. Dev made point that LW is used in so many scenarios they consider all of those, but NT are aware of the desire to see GPU in LW. I don't know what the practicalities would be of bolting Blender to your pipeline and consider Cycles till something more concrete happens.

    Good luck anyway.
    Personally, I take everything with a grain of salt. I guess I am really just looking for an unequivocal "yes" to the question of native GPU rendering, and I am less concerned with a timeline or estimate of any kind.

    If I knew there was an internal commitment to providing GPU rendering, it would impact the resources I expended now on an external solution, even without a timeline or estimate of any kind.

  6. #6
    ack ack Markc's Avatar
    Join Date
    Feb 2006
    Location
    England
    Posts
    1,370
    I am actually shocked (but pleased), there is 'rumor' of AMD support.
    Being a Mac user, this is good news if it happens.....
    Mac Pro (2010) OSX 10.13 RX580 Pulse LW2019

  7. #7
    I guess my question would be - what is slow? In my experience with Octane - which is freaky fast with even a moderate GPU - time is wasted setting up materials when comparing it to using native LW2019. I think it really depends on what kind of work you do. I do zero animations so all my hi-res requests render overnight and my day to day stuff (1200 x 1200) renders in the 1 to 2 minute range. That's plenty fast.
    Last edited by Tim Parsons; 02-22-2019 at 12:05 PM.
    Tim Parsons
    Technical Designer
    Sauder Woodworking Co.

    http://www.sauder.com

  8. #8
    ack ack Markc's Avatar
    Join Date
    Feb 2006
    Location
    England
    Posts
    1,370
    That is one thing that puts me off Octane/Keyshot etc, having to build surfaces for the render engine.
    Plus incompatibilities with third party LW plugins.

    Ideally LW needs a native gpu (option) within vpr.
    Mac Pro (2010) OSX 10.13 RX580 Pulse LW2019

  9. #9
    Registered User
    Join Date
    May 2012
    Location
    Virginia
    Posts
    428
    Quote Originally Posted by Tim Parsons View Post
    I guess my question would be - what is slow? In my experience with Octane - which is freaky fast with even a moderate GPU - time is wasted setting up materials when comparing it to using native LW2019. I think it really depends on what kind of work you do. I do zero animations so all my hi-res requests render overnight and my day to day stuff (1200 x 1200) renders in the 1 to 2 minute range. That's plenty fast.
    That very much describes our current process exactly, with the difference that we do a fair amount of animation. Your point about setting up materials in Octane is also one of my concerns.

    The problem is that previously we were compositing in OBJ sequences in After Effects using Element 3d and Trapcode for particle effects, etc.

    Now, we want to move that back to Lightwave to take advantage of some of the new tools and effects.

    Unfortunately, rendering a single frame of a similar animation is taking an excruciatingly long time using LW's native render. It looks great... but the time sink is not necessarily worth the results.

  10. #10
    Super Member
    Join Date
    Sep 2003
    Location
    Washington DC
    Posts
    3,627
    thanks for posting this question
    This message does not reflect the opinions of the US Government, CG Networks or CGTALK.com. The opinions expressed on this posting are on my own volition.

  11. #11
    Hi ! LW 2919 & Octane 4 are impressive !
    It's a viable solution for some pro works ! But ..

    01_It really must be remembered that the GPU rendering is still new ! Cards do not have many vram, drivers change all the time, don't forget pci lanes & airflow, power supply, .. The machine + software setup is very important to work ! You need backups of each version, validated in production, for drivers, software versions, plugins, .. Very difficult to do because you have to be permanently connected for the authentications, with all the auto updates to kick .. Winning case : CG update in GPU rendering .. And keep in mind that GTX / RTX cards are "gamer cards", Quadro are expensive, but pro .. BTW, take a good guarantee ! I burn 4 Titans, largely amortized, but never happen in CPU ..

    02 _ The rendering engine of LW will always be more precise, and will manage some FX (gradiant vol ligth, gradiant edge celshad, multiple trsp VBD, ..) that the GPU renderings can not yet manage .. It is important to master the 2 renderings ! Octane is just a renderer, what about if Juanjo stop this work ? Otoy will continue LW ?? The stability & the quality of the CPU rendering is the more important ! The render farm is accessible & compatible if needed ! In any case, you have to master the 2, but prepare to enter a world a little more "video games", a lot of R&D is needed, as well as bigggs hardware tests !

    But it's playable, I've been doing it for 5 years, ok, i love risks, we have only a life for rendering ! On the other hand a big BRAVO to the development team, the new LW 2019 is superb !!! An integration of Cycles can be interesting, but it will always be simplified .. Can't wait to see the new creations of users with the new CPU native engine ! LW powa !!!

    BTW, nice continuation to you ! And take the time to work on your workflow ! He will return it to you !
    _
    Stéphane AGULLO
    _________________

    FREE 3D \ VR MAKER
    _________________

    SA3D | G+| FB | SKF

  12. #12
    Registered User rdolishny's Avatar
    Join Date
    Jul 2003
    Location
    Peterborough, Ontario, Canada
    Posts
    313
    I love the look and of course speed of Redshift, but for every commercial project we have GarageFarm on standby and send all of our frames there to render. It's like having a personal render farm (well, really, it is) It's fast and accurate and we love it. Render times really don't mean anything to us.

    Now, with Redshift maybe I would change our workflow back to local rendering.

  13. #13
    Registered User
    Join Date
    Mar 2016
    Location
    Oxford, UK
    Posts
    794
    Quote Originally Posted by rdolishny View Post
    I love the look and of course speed of Redshift, but for every commercial project we have GarageFarm on standby and send all of our frames there to render. It's like having a personal render farm (well, really, it is) It's fast and accurate and we love it. Render times really don't mean anything to us.

    Now, with Redshift maybe I would change our workflow back to local rendering.
    Are you rendering Lightwave in Redshift springboarding via another software, or do you just use different softwares and renderers ?

    Re Ocatne, one thing I dislike is the way they introduced a turgid dongle license which still has to connect to the net every few months to update.
    Last edited by TheLexx; 02-22-2019 at 06:13 PM.

  14. #14
    Super Member CaptainMarlowe's Avatar
    Join Date
    May 2007
    Location
    Belfort, France
    Posts
    1,846
    Quote Originally Posted by Markc View Post
    I am actually shocked (but pleased), there is 'rumor' of AMD support.
    Being a Mac user, this is good news if it happens.....
    I filed a feature request for this with these reasons a while back. I guess the more people ask for it, the more it could be taken into consideration. So maybe you could file a freq too.

    - - - Updated - - -

    Quote Originally Posted by Markc View Post
    I am actually shocked (but pleased), there is 'rumor' of AMD support.
    Being a Mac user, this is good news if it happens.....
    I filed a feature request for this with these reasons a while back. I guess the more people ask for it, the more it could be taken into consideration. So maybe you could file a freq too.

  15. #15
    I've started using octane at 2014 and at first it was big PAIN to convert my arch-viz stuff to Octane materials (but then again i need to convert them for LW 2012019 also if i made them in 2015 and before) 'coz it requires all new surfacing, BUT with time it got easier/faster (there is also automatic convert automatic materials which is good but not perfect) and currently i'm using octane render for 99% of my LW projects. Thing is that if you have lot of premade models then it's gonna be long process BUT if you are starting each project as new separate thing (no need to reuse models/materials) then it's same process, you still need to surface them for native render also so surfacing time is no different if you are surfacing for octane form start, heck it's even faster with decent GPU you get really fast IPR results.

    BUT i'd argue that for nice octane usage you should start with 2 GPUs at least 'coz with single GPU you will need to use priority low since your GPU will choke screen/mouse movements when rendering and navigating views same time.

    Also going octane full speed might require upgrading your PSU and it is expensive so it just depends what are you doing. I often render interior animations and changing light conditions so native LW render only can do that in Montecarlo Bruteforce GI (no baking for changing light conditions and Interpolated is splotchy) so render-times without octane would be impossible for me to give my clients 30-40-50 seconds Interior animations in 2-3-4 days ready. BUT i do have 10 GPUs in 3 machines so it's "expensive" also.

    BUT best thing is that you have options, it's not like if you go to Octane route you must drop all other options, you still can render native LW if you wish/have better suited project for that. so you actually are adding more options for yourself.

    cheers
    RAM-Studio
    WS - Dual Xeon E5-2698v4/128GB/Win10x64/x TitanX(M) + 2x GTX 1080Ti + GTX 1080
    My LWM Video Car Modeling Tutorial

Page 1 of 6 123 ... LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •