This allow to copy entire texture in a faster way than using framebuffer
blitting.
This uses ARB_copy_image extension if available and fallback to
glCopyTexSubImage2D for older gl version.
Both method should be as fast if not faster than the framebuffer blitting.
```
...\gpu_texture.c(466,7): warning C4555: result of expression not used
...\gpu_texture.c(559,7): warning C4555: result of expression not used
...\gpu_texture.c:1205:72: warning: pointer targets in passing argument 4 of ‘glGetTexLevelParameteriv’ differ in signedness [-Wpointer-sign]
```
This commit adds a `mipmaps` member to the `GPUTexture` struct and also
computes to the memory used by these mipmaps and the memory used for
textures that are created from an external bindcode.
So it solves the following inconsistencies:
- The memory value for mipmaps was not being computed.
- As `GPU_texture_from_bindcode` didn't call
`gpu_texture_memory_footprint_add`, it brought inconsistencies to the
value of the used memory, especially when the texture is freed.
Differential Revision: https://developer.blender.org/D3554
We implement cubemap array support for EEVEE's lightcache reflection probes.
This removes stretched texels and bottom hemisphere seams artifacts caused
by the octahedral projection previously used.
This introduce versioning code for the lightcache which will discard any
lightcache version that is not compatible.
Differential Revision: https://developer.blender.org/D7066
This is more in line with standard grids and means we don't have to make
many special exceptions in the upcoming change for arbitrary number of volume
grids support in Eevee.
The workbench shader was also changed to fix bugs where squared density was
used, and the smoke color would affect the density so that black smoke would
be invisible. This can change the look of smoke in workbench significantly.
When using the color grid when smoke has a constant color, the color grid
will no longer be premultiplied by the density. If the color is constant
we want to be able not to store a grid at all. This breaks one test for
Cycles and Eevee, but the setup in that test using a color without density
does not make sense. It suffers from artifacts since the unpremultiplied
color grid by itself will not have smooth boundaries.
Differential Revision: https://developer.blender.org/D6951
Based on @fclem's suggestion in D6421, this commit implements support for
storing all tiles of a UDIM texture in a single 2D array texture on the GPU.
Previously, Eevee was binding one OpenGL texture per tile, quickly running
into hardware limits with nontrivial UDIM texture sets.
Workbench meanwhile had no UDIM support at all, as reusing the per-tile
approach would require splitting the mesh by tile as well as texture.
With this commit, both Workbench as well as Eevee now support huge numbers
of tiles, with the eventual limits being GPU memory and ultimately
GL_MAX_ARRAY_TEXTURE_LAYERS, which tends to be in the 1000s on modern GPUs.
Initially my plan was to have one array texture per unique size, but managing
the different textures and keeping everything consistent ended up being way
too complex.
Therefore, we now use a simpler version that allocates a texture that
is large enough to fit the largest tile and then packs all tiles into as many
layers as necessary.
As a result, each UDIM texture only binds two textures (one for the actual
images, one for metadata) regardless of how many tiles are used.
Note that this rolls back per-tile GPUTextures, meaning that we again have
per-Image GPUTextures like we did before the original UDIM commit,
but now with four instead of two types.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D6456
Adds a check when starting blender if your platform is supported. We use a blacklist
as drivers are updated more regular then blender (stable releases).
The mechanism detects if the support level changed or has been validated by the user previously.
Changes can happen due to users updating their drivers, but also when we change the support
level in our code base.
When the user has seen the limited support level message it is saved in the user config.
It would be better to have a system specific config section, but currently not clear
what could benefit from that.
When the platform is unsupported or has limited support a dialog box will appear including a link
to our user manual describing what to do.
**Windows**
Windows uses the MessageBox that is provided by the windows kernel.
**X11**
We use a very lowlevel messagebox for X11. It is very limited in use and can be fine tuned when needed.
**SDL/APPLE**
There is no implementation for SDL or APPLE at this moment as the platform support feature targets mostly Windows users.
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D5955
The AMD PRO driver on linux PROXY check also fails. Now the
configuration ATI/Unix/Official driver will also bypass the
Proxy test.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D5205
This reverts commit e061cb4437.
This function is problematic on some operating systems.
# Conflicts:
# source/blender/draw/intern/draw_manager.c
# source/blender/gpu/GPU_texture.h
# source/blender/gpu/intern/gpu_texture.c
Cycles now uses the color space on the image datablock, and uses OpenColorIO
to convert to scene linear as needed. Byte images do not take extra memory,
they are compressed in scene linear + sRGB transfer function which in common
cases is a no-op.
Eevee and workbench were changed to work similar. Float images are stored as
scene linear. Byte images are compressed as scene linear + sRGB and stored in
a GL_SRGB8_ALPHA8 texture. From the GLSL shader side this means they are read
as scene linear, simplifying the code and taking advantage of hardware support.
Further, OpenGL image textures are now all stored with premultiplied alpha.
Eevee texture sampling looks a little different now because interpolation
happens premultiplied and in scene linear space.
Overlays and grease pencil work in sRGB space so those now have an extra
conversion to sRGB after reading from image textures. This is not particularly
elegant but as long as engines use different conventions, one or the other
needs to do conversion.
This change breaks compatibility for cases where multiple image texture nodes
were using the same image with different color space node settings. However it
gives more predictable behavior for baking and texture painting if save, load
and image editing operations have a single color space to handle.
Differential Revision: https://developer.blender.org/D4807
In the case of the report a GL_PROXY_TEXTURE_2D_ARRAY of 2509x2509x1 failed to be allocated.
This is a work around as the GL_PROXY_TEXTURE_* is not reliable.
Reviewed By: brecht, fclem
Maniphest Tasks: T63223
Differential Revision: https://developer.blender.org/D4651
BF-admins agree to remove header information that isn't useful,
to reduce noise.
- BEGIN/END license blocks
Developers should add non license comments as separate comment blocks.
No need for separator text.
- Contributors
This is often invalid, outdated or misleading
especially when splitting files.
It's more useful to git-blame to find out who has developed the code.
See P901 for script to perform these edits.
Was generating INVALID_FRAMEBUFFER here instead of failled texture alloc.
Add safety asserts in gpu_texture.c and clamp minimum size to 1 inside
GPU_offscreen_create.
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
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.
Adding a workaround in this case: we blit the depth buffer instead of the
stencil buffer and use the copy as the texture. This is slower but at
least it should work.
When calling glBlitFramebuffer on most (if not all) mac that have a GPU
from the Radeon Pro series, the data is not properly copied and only a
subset of the pixels are correctly copied.
This only happens when blitting the depth buffer if the depth buffer is
GL_DEPTH24_STENCIL8.
Changing the depth buffer format to GPU_DEPTH32F_STENCIL8 fixes the issue
but only works if blitting the depth componnent. The stencil componnent
still provoke issues when being copied.
This is caused by a driver bug that prevent us from rendering to (or even
binding) a texture mip level that is below GL_TEXTURE_MAX_LEVEL of the
target texture. This is fine in most drivers (and legal AFAIK) but not on
thoses Intels HDXXX + Windows.
As a fix we just put GL_TEXTURE_MAX_LEVEL lower (which is illegal because
it is undefined behaviour), but in practice it works ok and does not
trigger any warnings or errors.
This commit fixes most of the problems encountered on these GPUs (T56668).
