I see a lot of people saying things like “I tried it but had issues and maybe I’ll just wait for the final release.”
But you know what? What you have today basically IS the final release with no remaining showstopper issues, so if it does not work for you today, then the final release probably is not going to work for you either!
If you know of any big problems you should report them as bugs NOW, not wait for some dreamed-of final release where the devs will have magically used their psychic powers to read your mind and figure out there might be problems that only you are seeing!
The only difference between what we have today and the fabled final release is THE BUGS YOU REPORT!
I wasn’t referring to you really, just trying to express that I have seen a lot of people say “I’ll wait for the final release” recently and issues only get fixed if people actually report them.
Support for ancient hardware (> 10 years old) has gotten to the point where it was starting to prevent the implementation of modern features, it was also becoming a massive burden to maintain. Cutting support for hardware that never got around to supporting OpenGL 3 or later allowed them to cut out a lot of cruft.
The reason for the maintenance nightmare was the fact they had to make code work for hardware that did not have driver support for many years. Bugs in those drivers would not be fixed, so workarounds and hacks had to be done that would build up in the code over time.
Still, 2.8 remains a lot more conservative with PC requirements than the big commercial solutions, the devs. are even going through the pain of supporting more recent Intel iGPU’s so you can work with 2.8 on a current generation budget build.
So far as a game pipeline mainly modeling and a few rendering and animations it’s seem rock solid on my side except the usual crash when dealing with heavy scene or model that come from sculpt app.
Keep in mind saving is atrocious on heavy scene and viewport performance is piss poor!
Unfortunately, it seems the devs. never got around to fixing the performance issues, despite the issues themselves having known solutions and despite them being on the pre-release roadmap. Clement’s batch shading patch has not been committed yet and I’m not sure if they decided to move any serious effort to 2.81 (so they can focus on making everything faster than 2.79 than matching it).
Blender 2.8 it’s already stable and production ready software ,for God sake in two or three weeks we will have the final release
Of course it is stable, but if they didn’t improve the edit mode performance in this two weeks this will be a shame
Seriously, guys. When you take a look to some low priority bugs, 2.8 simply does not look finished.
Like this one https://developer.blender.org/T66461 or this one https://developer.blender.org/T66457
If you add to that the lack of dynamic paint for weight groups or volume rendering not able to respect mesh shape.
Old 2.7 bugs and limitations + not completely cleaned 2.8 tools + EEVEE limitations. To me, those reasons are serious enough to avoid to encourage people to rush on 2.80.
I’ve been working exclusively with 2.80 for about a year now. On my Mac it runs really stable with maybe a crash once every 2 weeks and only when I try out combinations of -for example- hair, softbody and rigid objects. Or when I use UNDO a lot with big meshes. On my Linux machine it crashes more often, daily. GPU rendering freezes sometimes, but I’m not sure if it’s Blender or CUDA drivers. According to some guy who uses a different distro than me, I’m using the wrong Linux distro.
This could be due to the nature of volumes, and a lot of other things, being more difficult to do in the fast rasterized way for Eevee as opposed to the much slower, but extremely accurate pathtraced way in Cycles.
I guess there would be a way, but Clement would not have had the time to really focus on it if Eevee (at least an initial form of it) was to be ready this month in 2019. Do recall how long it took to find a working, yet fast, method for soft shadows. Also note the tons of little graphics and rendering bugs he needed to tackle in the last 6 months.
Like i have said for what i do it is working as intended but i know many problems are still affecting 2.8 but they don’t bother my type of workflow for now.
There are something like 1,300 outstanding issues in the 2.80 bug tracker, but most of it is stuff that doesn’t crash the program and has a workaround or just generally isn’t the end of the world. One could wait forever in order to try to fix all these smaller bugs, but most people likely won’t see a difference.
I’m sure more serious bugs will be found in 2.80 once it is released. That’s actually a big reason to push it out the door because many more people will start using it once it’s declared official and that will turn up any nasty bugs that have been missed.
So to the original question of “should I completely migrate from 2.79 to 2.80 when 2.80 is released?” the conservative answer for a professional 2.79b user is probably “no” if you are comfortable and productive with 2.79b and can live without Eevee and all the other new features for now. Let other more adventurous users and those who have less to lose blaze the trail, and follow along cautiously when the pioneers report back without too many arrows in the knee.
In addition, previous observations from the BF indicated that switching to a longer release cycle does not objectively reduce the total number of bugs in an official build (as reports have indicated). In part because it is less likely many will be reported.
I think the future may be something like the continuous release process that projects like Chrome or VS Code use. It’s actually what people have gotten used to in the last couple years with the 2.80 development process where you just download nightly builds as often as you like. We just need to start saving all the nightly builds and let people subscribe to “last night” “best of last week” “best of last month” etc. channels. Periodically a proven best-of-month build could be announced as an official version.
“AMD A6 processor” doesn’t really tell anything, there is like at least 22 “A6” models with 9 different generations just for laptops. Post full model naming along with rest of your system specifications.
However even oldest A6 gpu core should be able to support 3.3 OpenGL which is needed for 2.80
