Upbge 0.3.0 - GPU armature skinning

[BluePrintRandom]

Currently our skinning method is not very fast -

For AAA quality actors we will need something faster than blender built in armature deform.

for inspiration from armoury or direct port

https://github.com/armory3d/armory/blob/4ecec19ad3acd5f2e77b033c313b2eef47c4ae4a/blender/arm/material/make_skin.py

https://github.com/armory3d/armory/blob/3239c985943252bcb0680d29f14d966124c090d3/Shaders/std/skinning.glsl

[BluePrintRandom]

https://github.com/UPBGE/upbge/tree/upbge0.3-hard-skinning

[Liebranca]

i can try out a gpgpu mambo in opengl through transform feedback.

advantage: no need for the rasterizer to get involved early on way it does in some attempts ive seen. and then the code can be reused for different operations further down the road.

con: it means managing some additional buffer objects so mandatory rerun of an init block if the base mesh is modified or monkeyed with. but who adds a loop to a skinned mesh anyhoo.

if someone can explain to me how im supposed to slap a new lib into the solution then great. im terrible at writting cmakelists.

[lordloki]

Here you can see an example of library spindle added to intern directory: https://github.com/UPBGE/upbge/commit/adc26b29d0aabdb982a4a4d16607c7dea6b35cf0#diff-4a69625a7f0a11cc34b285178af671a0

[lordloki]

If you want, I can give you upbge/github access to do the work in a branch. This way i can help you easy with CMakeLists or whatever

[Liebranca]

that clears it up, thank you.

im already working on an unpublished repo cloned from ge_23a_old_bullet branch, and since im going to experiment heavily i feel more comfortable keeping it separate. but thanks for the offer.

[lordloki]

Closing the issue following the new UPBGE tracker rules but keep open the feature to check later.