Page 6 of 21 FirstFirst ... 4567816 ... LastLast
Results 76 to 90 of 304

Thread: Seriously NewTek... This is your best moment for LW2020

  1. #76
    Registered User
    Join Date
    Mar 2016
    Location
    Oxford, UK
    Posts
    937
    @50one, the cartoon Tyrannosaurus in X-Ray mode is cool, no ?
    That said, despite everything they do seem quietly committed.

  2. #77
    Registered User
    Join Date
    May 2012
    Location
    Virginia
    Posts
    530
    The reason that LW2020 is taking so long:

    1. LW2020 will be the first version of Lightwave that is a single application.
    2. Lightwave will include GPU rendering.


    That is my completely unfounded speculation on the wait for LW2020.

  3. #78
    Registered User Rayek's Avatar
    Join Date
    Feb 2006
    Location
    Vancouver, BC
    Posts
    1,568
    Quote Originally Posted by RPSchmidt View Post
    The reason that LW2020 is taking so long:

    1. LW2020 will be the first version of Lightwave that is a single application.
    2. Lightwave will include GPU rendering.

    Both would require a complete rewrite: (1) of Lightwave (Layout), (2) of the new 2012019 render code.

    Neither is possible within the short span of time since 2019's release.

    But who knows? The dev team might have been at work these past three years on adding GPU support. Very, very doubtful, though. Not that many good GPU render engine developers out there, and I haven't seen a sign that Newtek hired someone for that particular job.
    Win10 64 - i7 [email protected], p6t Deluxe v1, 48gb, Nvidia GTX 1080 8GB, Revodrive X2 240gb, e-mu 1820. Screens: 2 x Samsung s27a850ds 2560x1440, HP 1920x1200 in portrait mode

  4. #79
    Carbon fibre dongleŽ 50one's Avatar
    Join Date
    Jan 2011
    Location
    Honeycomb hideout
    Posts
    2,386
    Quote Originally Posted by TheLexx View Post
    @50one, the cartoon Tyrannosaurus in X-Ray mode is cool, no ?
    That said, despite everything they do seem quietly committed.
    I guess I'm too spoiled by some other free software

    Honestly, their marketing efforts are pretty bad, how many times I've seen a shoe model in Modo?!
    Sooner or later people start asking if this is shoe modelling app lol.

    Would be great if they would show some heavy polygonal models that really emphasize the quality of that viewport.
    Oh wait, they cannot, anything pass a certain threshold will render the app unstable and slow down the viewport.

  5. #80
    Super Member Chris S. (Fez)'s Avatar
    Join Date
    Feb 2003
    Location
    Virginia
    Posts
    3,029
    Integration and GPU rendering...righhhhht.

  6. #81
    Registered User
    Join Date
    May 2012
    Location
    Virginia
    Posts
    530
    Quote Originally Posted by Rayek View Post
    Both would require a complete rewrite: (1) of Lightwave (Layout), (2) of the new 2012019 render code.

    Neither is possible within the short span of time since 2019's release.

    But who knows? The dev team might have been at work these past three years on adding GPU support. Very, very doubtful, though. Not that many good GPU render engine developers out there, and I haven't seen a sign that Newtek hired someone for that particular job.
    Click image for larger version. 

Name:	why-so-serious-memegenerator-ne-why-so-serious-joker-why-53128475.jpg 
Views:	4 
Size:	164.7 KB 
ID:	147136

  7. #82
    Registered User Rayek's Avatar
    Join Date
    Feb 2006
    Location
    Vancouver, BC
    Posts
    1,568
    Quote Originally Posted by RPSchmidt View Post
    Click image for larger version. 

Name:	why-so-serious-memegenerator-ne-why-so-serious-joker-why-53128475.jpg 
Views:	4 
Size:	164.7 KB 
ID:	147136
    Ah. The original post was meant to be sarcastic in the first place. Text-only comms remove subtle intent, unfortunately, so I did not catch that.
    Win10 64 - i7 [email protected], p6t Deluxe v1, 48gb, Nvidia GTX 1080 8GB, Revodrive X2 240gb, e-mu 1820. Screens: 2 x Samsung s27a850ds 2560x1440, HP 1920x1200 in portrait mode

  8. #83
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    6,877
    Quote Originally Posted by Rayek View Post
    Not that many good GPU render engine developers out there, and I haven't seen a sign that Newtek hired someone for that particular job.
    IMO, they'd really need both a GPGPU specialist (to re-architect the engine for GPU memory architecture and other differences), and a embedded/system-sw specialist to provide the missing code infrastructure services. Either by themselves isn't really enough, and you're talking at least 2+ engineer/years of work for each before they'd even have something basic processing end-to-end.

    Even most devs grossly underestimate the complexity and volume of "data logistics" needed to ensure all required data are available in the proper memory scope at the proper time, to satisfy both GPU memory architecture limitations, as well as all the different needs of an render engine. Just building, coordinating, and debugging the infrastructure to deal with those logistics will likely require most of an engineer-year for both GPGPU specialist and systems specialist. Embracing GPU tiered (and highly-constrained) memory handling is critical in obtaining "GPU-level performance".
    Last edited by jwiede; 03-24-2020 at 01:33 PM.
    John W.
    LW2015.3UB/2019.1.5 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

  9. #84

    Exactly so.
    do we know this?

    didn't 2018 get an engine rewrite?
    LW vidz   DPont donate   LightWiki   RHiggit   IKBooster   My vidz

  10. #85
    Registered User
    Join Date
    May 2012
    Location
    Virginia
    Posts
    530
    Quote Originally Posted by Rayek View Post
    Ah. The original post was meant to be sarcastic in the first place. Text-only comms remove subtle intent, unfortunately, so I did not catch that.
    Not sarcasm, as much as a bit of tongue in cheek.

  11. #86
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    6,877
    Quote Originally Posted by erikals View Post

    do we know this?

    didn't 2018 get an engine rewrite?
    Layout 2018 received a geometry database rewrite (Hydra), as well as a (larger) surfacing / render engine rewrite. The surfacing/rendering rewrite did little to enable modeling in Layout whatsoever. The Hydra rewrite did some, but by itself, wasn't adequate to provide everything needed to properly support even basic modeling in Layout. There's a ton more missing infrastructure needed, adding which would completely change the operation of Layout's main event loop and viewport UX handling, among other major impacts.

    Just as an example, consider what all has to be present to support the input-handling interaction for geo-entity-level (read as, point/edge/poly) snapping and selection in the viewport (and to maintain that across potential changes in geo-entity db contents). Even if we presume Hydra supplies both spatially-indexed as well as object-indexed access to arbitrary geo-entities, something still needs to be continually identifying "nearby candidate" entities based on raycasting viewport mouse positions (and "nearby" is a fairly complex proposition, even given a "spatially-indexed" geo db -- just consider how hidden geo, layers, etc. factor into that).

    You're taking Layout from spatially-casting and picking from a list of up to maybe a thousand potential objects (though Layout seriously bogs down moving towards a thousand objects), to spatially-casting and identifying from sets of up to millions, even tens-of-millions of individual geometry entities, with the requirement that it continue to do so in a usefully-UX-performant manner. And that's just to grant ability to be able to click on (or near) points/edges/polys and identify them as the target -- we haven't even covered integration with existing object-level functionality, nor managing geo-entity selections or so forth.

    Oh, and for grins, don't lose sight of the fact that Modeler only ever dealt with a single object at once, while a modeling-capable Layout will necessarily have to deal with ALL the geo, from ALL the objects in scene, at once. That's a huge total number of entities present where Layout will need to maintain "UX-efficient" operation. All from a team that historically hasn't exhibited particular strength in either performance or resource-footprint optimization (note: "not exhibited strength" != "weakness", just "not exhibited strength").

    Another critical question is whether it even makes sense to rewrite all that input / event loop and viewport UX code using LW's UI toolkit, or bail and instead rewrite it using Qt. Reimplementing it all using Qt means even more to do short-term, but gives much greater long-term capabilities. In contrast, as LW UI toolkit is essentially a dead-end, rewriting it for LW UI toolkit might mean less work short-term, but just adds more code needing reimplementation for the eventual, unavoidable Qt migration.

    Getting Hydra integrated was a step on the road to making Layout capable of modeling, but doing so clearly isn't "all" or even "most" of the effort/rewrite required to do so, there's still a LOT more needed they haven't had time to do (there's also question whether they have dev staff efficiently able to do so, as modeling position apparently wasn't filled).
    Last edited by jwiede; 03-24-2020 at 04:31 PM.
    John W.
    LW2015.3UB/2019.1.5 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

  12. #87
    Yep- all that is never going to happen.

    I think their only choice is to improve Modeler with the resources they have - I mean after all they had that big survey a while back about what we wanted to see improved in Modeler.
    Last edited by Tim Parsons; 03-24-2020 at 06:03 PM.
    Tim Parsons
    Technical Designer
    Sauder Woodworking Co.

    http://www.sauder.com

  13. #88
    Registered User
    Join Date
    Jun 2014
    Location
    Right here
    Posts
    1,803
    Quote Originally Posted by Tim Parsons View Post
    Yep- all that is never going to happen.

    I think their only choice is to improve Modeler with the resources they have - I mean after all they had that big survey a while back about what we wanted to see improved in Modeler.
    I agree. But improving Modeler is for no use in my opinion (except for longtime LW-only users). Every other 3D app is at least 10+ years ahead of Modeler. An improved Modeler will not attract any new user.
    .

  14. #89
    Registered User
    Join Date
    Aug 2016
    Location
    a place
    Posts
    2,273
    why not improve the hub instead and make is seamlessly integrate with 3rd party tools & applications?

    seems a better option that would provide infinite expansion of workflow.

  15. #90
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    6,877
    Quote Originally Posted by gar26lw View Post
    why not improve the hub instead and make is seamlessly integrate with 3rd party tools & applications?

    seems a better option that would provide infinite expansion of workflow.
    It's been requested for ages (expose Hub interchange protocol), we need it (plus Sentinal info) for another big reason as well (ensuring firewalls aren't causing issues), but requests never got any traction.

    That said, as it's a shared memory buffer-based protocol, it might be very tightly-coupled to Layout:Modeler internals. For example, it might be making a lot of presumptions about execution contexts, and so forth. If so, that'd make it less useful for use with anything else, though it'd still be valuable to have it formally described (for debugging and to identify whether certain recurring install/config problems' origins).
    Last edited by jwiede; 03-25-2020 at 08:29 PM.
    John W.
    LW2015.3UB/2019.1.5 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

Page 6 of 21 FirstFirst ... 4567816 ... 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
  •