Page 3 of 13 FirstFirst 12345 ... LastLast
Results 31 to 45 of 190

Thread: Discussion for Upcoming Performance Optimizations in The Next LightWave Release Blog

  1. #31
    Quote Originally Posted by robpowers3d View Post
    Very glad to hear that! ChronoSculpt is like my baby and I love it so much. It makes me happy to hear how it working so well for you. We also have a studio that is currently working to implement ChronoSculpt as a key part of their pipeline because of the unique workflow that it provides for editing dynamics simulations and editing cached mesh deformations over time with large data sets. It is a very useful tool for that type of workflow. Thanks for sharing your experiences.
    Rob
    Is there going to be an update to ChronoSculpt or will it be integrated into Lightwave NEXT?
    Will there be any audio integration for lip syncing setups?

  2. #32
    Registered User Norka's Avatar
    Join Date
    Nov 2015
    Location
    NE Ohio
    Posts
    236
    Will we be seeing a significant bump in performance when doing millions of instances?

  3. #33
    Hardcore Lightwaver Photogram's Avatar
    Join Date
    Sep 2010
    Location
    Montreal
    Posts
    209
    The new performances are very welcome!
    I like the new open GL shading, now parts of the model like inside the mouth are occluded and not lit as in previous open GL.
    I also note the lower memory usage compare to the 2015 version, that's another good thing!
    Playback speed is so important and these optimisation are a step in the right direction!
    I'm curious to see if the new mesh system will also going to be integrated in the Modeler
    Bravo! LW3DG

  4. #34
    Lighthearted Kaptive's Avatar
    Join Date
    Jan 2006
    Location
    Derby, UK
    Posts
    698
    Quote Originally Posted by Norka View Post
    Will we be seeing a significant bump in performance when doing millions of instances?
    please please please please

    It's gotta help with this, it must. LW is pretty good with this anyway (considering how intensive it can get), but it can always be better. Any speed ups are good speed ups.

  5. #35
    Quote Originally Posted by hrgiger View Post
    I thought that too about playing both at once but as was pointed out to me, playback is likely only using a single processor so it is doubtful that playing them seperately would make much difference.
    Hi,
    I hope the new Unified Mesh System use all processors for deformations...
    ciao
    Thomas
    Web: www.dieleinwandhelden.com

    I use two pieces of the three-piece application with mocap module.

  6. #36
    Super Member
    Join Date
    Sep 2010
    Location
    UK
    Posts
    383
    Thanks Rob for info on speed enhancements and great to see a demo video by Lino
    that shows the level of speed increase.

    Question to Lino.
    What computer system did you have Lightave NEXT running on at the time you produced the video [eg. cpu/graphics card]?.
    Also given that the figure is animating in perspective view will upgradng to a better graphics card improve the speed performance
    even further as i understand that upgrading your graphics card would normally give you better rotation etc of the scene when you
    have a complex scene [triangles/polygon rate per second].
    What im not shore about is wether a better graphics card would improve the speed of animations in perspective mode.

  7. #37
    Registered User
    Join Date
    Oct 2011
    Location
    Utah
    Posts
    365
    I am excited to get my hands on the new Lightwave. Thanks for the sneak peek!

  8. #38
    Super Member spherical's Avatar
    Join Date
    Dec 2004
    Location
    San Juan Island
    Posts
    4,686
    Speed of viewport draw is enhanced by using a Professional Level CAD card like a Quadro. Consumer cards don't do as well, but a faster card is always better; just not the speed improvement you're hoping for.
    Blown Glass · Carbon Fiber + Imagination

    Spherical Magic | We Build Cool Stuff!

    "When a man loves cats, I am his friend and comrade, without further introduction." - Mark Twain

  9. #39
    Member tburbage's Avatar
    Join Date
    Feb 2003
    Location
    Redwood City, CA
    Posts
    289
    Quote Originally Posted by robpowers3d View Post
    Very glad to hear that! ChronoSculpt is like my baby and I love it so much. It makes me happy to hear how it working so well for you. We also have a studio that is currently working to implement ChronoSculpt as a key part of their pipeline because of the unique workflow that it provides for editing dynamics simulations and editing cached mesh deformations over time with large data sets. It is a very useful tool for that type of workflow. Thanks for sharing your experiences.
    Rob, I hope you can tell that story at some point. That would be a good read!

    Really looking forward to seeing LW+NEXT. I'm ready for that PBR shading and rendering support.

  10. #40
    Scene Destroyer DrStrik9's Avatar
    Join Date
    May 2007
    Location
    Portland, Oregon
    Posts
    1,671
    Quote Originally Posted by tburbage View Post
    Rob, I hope you can tell that story at some point. That would be a good read!

    Really looking forward to seeing LW+NEXT. I'm ready for that PBR shading and rendering support.
    Agreed on the good read idea! (And the next idea too.) :+) I'd also love to hear the down-low nitty-gritty on the decision to graciously give two super-cool apps to current LW users, ... etc.

    I'm ready for the potentially rocky road of Next, when it appears.

    I do greatly appreciate the work of the visionaries and coders at LW3DG, and the poor customer service people who process bug-reports and take my angst and always give me peace in return.

  11. #41
    Registered User Schwyhart's Avatar
    Join Date
    Mar 2013
    Location
    Tulsa OK
    Posts
    291
    I really hope that polygon (face) selections appear as a solid selection instead of an outline. Please say it's so!
    I know I'm probably the only one who cares about that.

  12. #42
    Super Member Snosrap's Avatar
    Join Date
    Aug 2004
    Location
    Ohio, USA
    Posts
    4,611
    Quote Originally Posted by Schwyhart View Post
    I really hope that polygon (face) selections appear as a solid selection instead of an outline. Please say it's so!
    I know I'm probably the only one who cares about that.
    Doubt it. Looks like Layout is getting most the love again.

  13. #43
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    5,200
    Rob, I have a few questions about the conditions of the comparison:

    1. Was the same machine used for capturing both 2015 and LW.Next usage? What are the machine specs used (incl. drive(s) and interface used for MDD)?

    2. Can you describe how the captures were done? Live screencaptures or via camera? Were they done simultaneously or as separate runs (and why)?

    3. (edit) In the case of the MDD captures, what (if anything) was done to ensure equivalent disk caching, etc. state for both apps?

    4. Why no stopwatch/timecode embedded and shown in captured video? Would have been nice to have reliable basis for quantified comparison.


    Thanks!

    (edit) hrgiger seems to indicate screencapture of both apps were done simultaneously on same machine? I'm really hoping that wasn't the case, but if so, what precautions were taking w.r.t. CPU affinities, gfx scheduling, etc. to ensure "fair" comparison?
    Last edited by jwiede; 03-27-2016 at 01:18 PM.
    John W.
    LW2015.3UB on MacPro(12C/24T/10.12.5),32GB RAM, NV 980ti

  14. #44
    Well you can see the mouse move from one Layout viewport to the next so it seems clear he has both Layout 2015 and Layout beta version running at the same time on the same machine. I guess I just wonder why the video wouldn't be a fair comparison if they are running on the same machine simultaneously?
    Last edited by hrgiger; 03-27-2016 at 01:45 PM.

  15. #45
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    5,200
    Quote Originally Posted by hrgiger View Post
    I thought that too about playing both at once but as was pointed out to me, playback is likely only using a single processor so it is doubtful that playing them seperately would make much difference.
    Untrue, doing so will result in huge, comparison-invalidating differences depending on how memory and gfx resources are mapped/cached, how much physical memory each has available, etc. Running both simultaneously won't even reliably guarantee all their threads wind up on different CPUs, unless explicit steps are taken ahead of time. Foreground vs background UI state also explicitly impacts scheduling for both CPU and GPU, resource priorities, esp. in comparisons of the type shown (there are some visible indications LW2015 was at a disadvantage in that regard).

    The reason why such perf. comparisons are done as discrete, foreground app runs (typically after n "warming" runs) in controlled, clean-setup environments is because that's the only approach which ensures both apps receive the same advantages at the same levels during execution. Without that equalization, such environmental factors drive the margin of error greater than the actual perf. differences measured, totally invalidating the comparison.

    - - - Updated - - -

    Quote Originally Posted by hrgiger View Post
    Well you can see the mouse move from one Layout viewport to the next so it seems clear he has both Layout 2015 and Layout beta version running at the same time on the same machine. I guess I just wonder why the video wouldn't be a fair comparison if they are running on the same machine simultaneously?
    Foreground vs background scheduling differences alone are typically sufficient to invalidate such comparisons. Even if you equalize that, you also need to balance process/thread affinity mappings, CPU and GPU working set allocations, and resource caching before there's any reason to think relative performance is usefully comparable. It's a huge effort required to level the playing field for simultaneous comparison, and for complex apps generally necessitates virtualization or special drivers.

    Ultimately, it's MUCH more efficient and reliable to instead ensure isolated app measurement runs incurring equivalent advantages and disadvantages, which is why that's the approach generally used (unless there's an absolute need for simultaneous comparison, which is extremely rare).
    Last edited by jwiede; 03-27-2016 at 02:43 PM.
    John W.
    LW2015.3UB on MacPro(12C/24T/10.12.5),32GB RAM, NV 980ti

Page 3 of 13 FirstFirst 12345 ... 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
  •