While it's one additional indirection (that has an extra cost on Intel
GPUs apparently, like with Phong and MeshVisualizer and
DistanceFieldVector already), with the assumption that draws usually
share the material info it allows to cram more draws into the 16/64k UBO
limit as the per-draw data are now one vec4 smaller.
For the indirection overhead I can imagine adding a new flag which makes
material mapping implicit (materialId == drawId). That seems to put the
benchmark numbers back to the original speed. Same could be done for
other shaders.
No issue on GCC, only Clang and MSVC started failing once I renamed the
shaders. These four are the only that don't have the export macro
attached AND YET it somehow worked without any problem until now.
This was stupid, eh? Blame Mesa and SwiftShader for not exposing
ANGLE_instanced_arrays so the only way to test this for me was via the
browser, which is practically impossible. Then found this by an
accident.
Pushing straight to master because YOLO.
Except MeshVisualizer and VertexColor, which don't have any texturing,
so there it's not needed. In most cases the tests are reusing existing
ground truth files and only modifying transformations / flipping images.
Bloaty says it saved 10 kB in Debug build of MagnumGL:
VM SIZE FILE SIZE
-------------- --------------
[ = ] 0 .debug_info +1.59Ki +0.0%
+0.4% +1.50Ki .text +1.50Ki +0.4%
[ = ] 0 .debug_str +409 +0.0%
[ = ] 0 .debug_line +276 +0.1%
[ = ] 0 .debug_abbrev +20 +0.0%
-28.6% -2 [LOAD [RX]] -2 -28.6%
[ = ] 0 [Unmapped] -4.28Ki -41.0%
-22.7% -9.23Ki .rodata -9.23Ki -22.7%
-0.8% -7.73Ki TOTAL -9.73Ki -0.1%
And 4 kB in Release:
VM SIZE FILE SIZE
-------------- --------------
+1.1% +3.44Ki .text +3.44Ki +1.1%
+1.7% +1.39Ki .eh_frame +1.39Ki +1.7%
[ = ] 0 [Unmapped] +656 +51%
-25.5% -9.47Ki .rodata -9.47Ki -25.5%
-0.7% -4.64Ki TOTAL -4.00Ki -0.4%
That's not negative, so I guess that's good. This change is of course
more significant in the context of a minimal WebGL build, where the exe
can be as little as 50 kB -- there 4 kB is almost 10% of the size.
This only generates code that will be never executed. Tested with Flat
and Phong, the other shaders don't have rendering tests yet but since
the change is the same, I assume it will work there as well.
What's left is *a lot* of places taking monstrous
std::vector<std::reference_wrapper> and that can't be changed to
std::vector<Containers::Reference> in a source-compatible way. Even that
would be only a temporary change, since the goal is to fully avoid
dependency on STL in those cases.
The final version of these APIs should take
Containers::ArrayView<Containers::Reference> and be implicitly
convertible froom e.g. std::vector<Containers::Reference>. That's
definitely possible, but not in time for 2019.01, so instead of forcing
users to temporary pass a `{vec.begin(), vec.size()}` everywhere instead
of just `vec`, I'm rather keeping these APIs intact.
Slow and ugly, is here only for making quick'n'dirty alpha masked
drawing without a need for blending or depth sorting. Oh and also to
support the glTF alpha mask feature. Again, beware: *slow*.
This better reflects that the functions modify a global state instead of
a shader-local state and so rebinding may be necessary (unlike with
uniforms, which get preserved).
The old set*() functions are now inline aliases to the bind*()
functions, are marked as deprecated and will be removed in some future
release.
103% of use cases use the returned value directly without checking, so
we might as well do the check ourselves. Added new function hasCurrent()
and added deprecated backward-compatibility conversion and -> operators.
Wow, that creeped to a lot of places.
Last dinosaur from the pointer age.
Also probably fixed a few issues when compiling the shader on older GLSL
and GLSL ES (floating point literal suffixed, missing precision qualifiers).
And less crazy preprocessor.
Each class/function that needs to access the resources first checks
whether the group exists and the group is registered if not. Thus there
is now no difference and annoying special cases when using static build.
Compiling fragment and vertex shader simultaenously, at least. Nothing
more can be done for now.
Also removed weird duplicate compile/link calls from MeshVisualizer,
went unnoticed since b9a72bd3d1. Why did I
do that?!
Why did I do this:
* It is more clean, shorter and nice looking with method chaining,
i.e. instead of:
shader.setColor(...)
.setOtherParam(5);
texture1.bind(MyShader::Texture1Layer);
texture2.bind(MyShader::Texture2Layer);
We now have this:
shader.setColor(...)
.setOtherParam(5)
.setTexture1(texture1)
.setTexture2(texture2);
* It is now also clear which texture type is expected, the layer
constant did not say anything about type.
* Also it is possible to use new features (multi bind, bindless
textures etc.) while preserving the same public API.
The only potential disadvantage is that the textures don't stay bound
like uniform values do, but this become a non-issue with bindless
textures. As usual, the old way is now deprecated and will be removed in
some future release.
The only places where they aren't absolute are:
- when header is included from corresponding source file
- when including headers which are not part of final installation (e.g.
test-specific configuration, headers from Implementation/)
Everything what was in src/ is now in src/Corrade, everything from
src/Plugins is now in src/MagnumPlugins, everything from external/ is in
src/MagnumExternal. Added new CMakeLists.txt file and updated the other
ones for the moves, no other change was made. If MAGNUM_BUILD_DEPRECATED
is set, everything compiles and installs like previously except for the
plugins, which are now in MagnumPlugins and not in Magnum/Plugins.
There will be many places (e.g. all
Platform::*Application::Configuration classes) where Version will be
used without Context (and all GL stuff brought with it).
Vladimír Vondruš