It might become necessary to take a look at that to stay in competition though - with modo having the self proclaimed fastest modeling toolset and autodesk expressing plans of bringing maya editing performance up to par with the one in 3dsmax it seems something that’s needed to stay in competition.
Good to know though, I need to lower my expectations regarding that ticket about performance regressions.
I think you are worrying too much. The core team will probably turn their attention to this within the 2.8 series. They just can’t do everything at once.
Blender is open source, so even if the core team doesn’t work on it any time soon, an outside developer could show up and improve that for free. This is exactly why blender is so powerful despite only getting 5500 euros each month in donations for a very long time, and since they get 35k atm they are improving blender at a faster rate than before.
It’s not that I think it’s never going to be done, it’s more about having to buy/renew some susbcriptions until then
But yes, I agree. Also, at least with the open source nature of Blender there is a level of transparency during the development which is really helpful when planning.
At the moment this is just a speculation, but on this video they say that in the future multitreading will be automatically managed by some predictive machine learning algorithm that will sort the various treading via hardware making the best use of resources.
So the developers will continue to create linear code and the hardware will take care of sorting the resources between the various cores.
I’ve got a scene with 1.3 million polys, and there’s a definite slowdown when moving anything in the scene in 2.8 (latest builds too). Even just rotating the camera around in solid mode is incredibly laggy. In 2.79 it’s as buttery smooth as you could ask for though. Hopefully they can get this ironed out before release.
those high priority bugs and performance reggressions will be fixed in middle june
according to code.blender.org
< * Mid June: fix all high priority bugs. Other bugs may be fixed as well, but bugs marked as high priority are the ones that we consider to block the final release.>
so right now there is a 58 high priority task, not so many !! https://developer.blender.org/maniphest/query/LVuBeJukZpzX/
Sergey Sharybin “has decided to take the bull by the horns” … he is investigating the real reasons why super heavy meshes become difficult to edit https://developer.blender.org/T64261
May he be successful in his journey through the forgotten depths of Blender’s guts, and may we see him get out unscathed, brandishing the trophy of clever optimization, and making this world a better world for all of us
Clément Foucault’s changes today appear to have greatly improved the slow object selection issue when a scene has thousands of particle/mesh instances etc. at least in some cases. So if this is a problem you’ve encountered, give tonight’s build a try tomorrow
For me on a specific scene where it used to take about six seconds to get the selection highlight to move from one object to another, which involved 100% GPU utilization that went on for several seconds after the highlight switch, it’s now (almost) instant and there’s only brief GPU utilization.
Any update on this since i recently start working on my heavier models that work fine when i was working with 3ds max but in Blender latest version edit mode is sluggish and barely respond while in 3ds max it was buttery smooth?
I think this is actually what Clément was working on when he made the changes that happened to improve the selection issue, so they’re probably still working on the problem of really slow edit mode operations on high-poly models.
The 2.8 development plan has the devs. working on blocking issues (which includes slow editmode especially with subsurf) in mid June. This will supposedly be one of the last steps taken before the big release.
The point is to get editmode and deformation performance to the level of 2.79, 2.81 and onward will probably see further improvements.