This replaces the test of consistency and capacity made with `GL_PROXY_TEXTURE_..` on AMD GPUs with one that checks only if the texture fits the limits of size and layer.
Differential Revision: https://developer.blender.org/D4081
This is a small change. We delay all gl calls at the first use of the
GPUIndexBuf / GPUVertBuf in order to be able to create multiple buffers
from different threads without having many gl contexts.
Textures are now hooked up to the RESET operation of particle
settings, which ensures particles being re-distributed when
texture is changed.
This is limited to a direct user modifications, which matches
old behavior in 2.79.
Right now does not add padding at the end of the buffer.
This seems not necessary but may cause problem on some platform. If needed
we will add this padding (only 2 more vertices).
We exploit the fact that we are using the metallic workflow for material
and pass the metallic parameter instead of the specular color.
Pack the front facing bit in the color buffer only for matcap display.
Change buffer formats to use less bytes as possible.
Also don't request buffers that we won't use.
Saved 40MB on 2K screen on StudioLight + Shadows + Specular Lighting.
Includes several cleanups.
Instead of doing a lot of alpha blended drawing with jittering, use the
fragment shader to do the masking using a circle mask.
This is much simpler and requires much less resources.
Hopefully this may solve the issue we have with the Intels UHD Graphics 620
on linux.
That was caused by a thread safety issue on gpu_batch_presets_unregister()
which was not designed to be used for this kind of situation (managing 3D
meshes batches).
This is a nasty bug. Because the node does not get properlly tagged as SSS
(sss_id is 0) but is still evaluated (so tagging the GPUMaterial as having
SSS). The sssProfile UBO is still declared and we need to bind something
to it.
On some platform does not support line width > 1.0 and can even throw and
error. Better check an at least display something rather than no lines at
all.
Node group inputs should behave like cycles now.
----
We create dummy nodes that contains the default values for the nodegroup
inputs and link them to the sockets. This way the uniform buffer gathering
function can read them.
But we also need to disconnect all the internal links to sockets that have
hidden values. Theses sockets must not get the values from the nodegroup
input sockets if there is no link to them otherwise we get broken results
(i.e.: normals for a bsdf node).
