This makes incremental recompilation faster when changing static
interned strings.
Differential Revision: https://code.wildfiregames.com/D3975
This was SVN commit r25457.
The 'arena' allocator does not take memory alignment into account, which
can result in crashes at -O3 with gcc 7.5 (presumably because of SSE
instructions).
This accounts for alignment issues, fixing the issue.
Also do various cleanup in lib/allocators.
Reported by: Bellaz89, Freagarach
Comments by: vladislavbelov
Differential Revision: https://code.wildfiregames.com/D3181
This was SVN commit r24517.
- Rename macros to be more explicit
- Move detection code to a separate file
- Remove a lot of checks in ARB mode (ModelDef.cpp would check for sse
multiple times per frame)
- Make explicit the SSE2 dependency for Windows
Comments by: @vladislavbelov@wraitii@OptimusShepard
Differential Revision: https://code.wildfiregames.com/D3212
This was SVN commit r24489.
ShaderModelRendererInternals is defined twice, once by ModelRenderer.cpp
and once by HWLightingModelRenderer.cpp.
Having two different definitions in the global namespace is a violation
of the C++ One-Definition-Rule.
Patch by: StefanBruens
Reviewed By: wraitii, Vladislav
Differential Revision: https://code.wildfiregames.com/D2932
This was SVN commit r23941.
Replace boost::hash_combine with a lib/hash.h hash_combine performing
the same statement.
Replace inconspicuous global boost hash_value specializations with
std::hash specializations.
No performance difference was observed in three simple MeshManager
measurements.
Remove unused TAG_MASK and h_tag in h_mgr.cpp following 0748c5a75e.
Replace typedef with using keyword and sort header includes.
Differential Revision: https://code.wildfiregames.com/D2441
Tested on: clang 9.0.0, gcc 9.2.0, Jenkins/vs2015, Jenkins/gcc6
This was SVN commit r23191.
- Move the * and & to the correct side,
- Add .0 and 0.f to clearly mark the types
- Pass pointers instead of arrays
- Add a newline after forward declaration
- Add spaces between operators
- Use c++ cast
Reviewed by: @vladislavbelov
Differential Revision: https://code.wildfiregames.com/D1934
This was SVN commit r22323.
Everything is char* now, so we don't need to mess around with different
string types.
Done with:
ag -ls 'LOG(MESSAGE|MESSAGERENDER|WARNING|ERROR)' source | xargs perl
-pi -e'1 while
s/(LOG(MESSAGE|MESSAGERENDER|WARNING|ERROR).*)%[hl]s/$1%s/g'
This was SVN commit r16187.
Done with:
ag -ls 'LOG(MESSAGE|MESSAGERENDER|WARNING|ERROR)' source | xargs sed
-i 's/LOG\(MESSAGE\|MESSAGERENDER\|WARNING\|ERROR\)(L/LOG\1(/g'
This was SVN commit r16183.
Previously we had a single culling frustum based on the main camera, and
any object outside the frustum would never get rendered, even if it
should actually contribute to shadows or reflections/refractions. This
caused ugly pop-in effects in the shadows and reflections while
scrolling.
Extend the renderer to support multiple cull groups, each with a
separate frustum and with separate lists of submitted objects, so that
shadows and reflections will render the correctly culled sets of
objects.
Update the shadow map generation to compute the (hopefully) correct
bounds and matrices for this new scheme.
Include terrain patches in the shadow bounds, so hills can cast shadows
correctly.
Remove the code that tried to render objects slightly outside the camera
frustum in order to reduce the pop-in effect, since that was a
workaround for the lack of a proper fix.
Remove the model/patch filtering code, which was used to cull objects
that were in the normal camera frustum but should be excluded from
reflections/refractions, since that's redundant now too.
Inline DistanceToPlane to save a few hundred usecs per frame inside
CCmpUnitRenderer::RenderSubmit.
Fixes#504, #579.
This was SVN commit r15445.
Use an arena allocator in ShaderModelRenderer::Render, to reduce
the allocation cost in STL containers.
Avoid unnecessary copying of std::vectors.
This was SVN commit r13911.
Switch all the constant strings in graphics code to use the new
variables.
This avoids the cost of instantiating CStrInterns at runtime every
frame.
This was SVN commit r13906.
CShaderDefines is designed to be efficient to copy and compare, but not
to construct. Conditional defines were constructing new CShaderDefines
for many models every frame.
Precompute all the possible conditional combinations of CShaderDefines
when a material is first loaded, so they can be looked up at no cost
when rendering.
This was SVN commit r13902.
