* TileLayer: use shared_ptr for vector framespec
This reduce memory copy of TileLayer from (4 to 16) * FrameSpec where FrameSpec = (sizeof(int) + 3 * sizeof(ptr)) to int + sizeof(ptr)
Callgrind difference
Before: https://lut.im/RGkiJqQb8T/LeQIEXpAuRzfl7gd.png
After: https://lut.im/bcqmwee1xu/cTwtptY5tRuS9lp0.png
* Fix one push_back to use vector::emplace_back & optimize inclusions
* VoxelManip cleanups (const ref, const move) permitting to improve a little bit performance
* VoxelArea: precalculate extent (performance enhancement)
This permits to reduce extend high cost to zero and drop many v3s16 object creation/removal to calculate extent
It rebalance the client thread update to updateFastFaceRow instead of MapBlockMesh generation
This will also benefits to mapgen
* Add Device3D class which will contain IrrlichtDevice interface
move getSupportedVideoDrivers to Device3D
Add Device3D singleton & use it in various places
Rename Device3D to Rendering engine & add helper functions to various device pointers
More singleton work
RenderingEngine owns draw_load_screen
move draw functions to RenderingEngine
Reduce IrrlichtDevice exposure and guienvironment
RenderingEngine: Expose get_timer_time() to remove device from guiEngine
Make irrlichtdevice & scene manager less exposed
* Code style fixes
* Move porting::getVideoDriverName, getVideoDriverFriendlyName, getDisplayDensity, getDisplaySize to RenderingEngine
Fix XORG_USED macro -> RenderingEngine + create_engine_device from RenderingEngine constructor directly
* enum paralax => enum parallax
* C++11 patchset 10: continue cleanup on constructors
* Drop obsolete bool MainMenuData::enable_public (setting is called with cURL in server loop)
* More classes cleanup
* More classes cleanup + change NULL tests to boolean tests
thread_local permits to limit variable lifetime to thread duration. Use it on each setting place which uses static to cache variable result only for thread lifetime. This permits to keep the same performance level & reconfigure server from MT gui in those various variables places.
Add thread_local to undersampling calculation too.
* Optimize updateFastFaceRow processing by removing some TileSpec copy
It permit to decrease this function from 54% runtime to 45% and reduce copy from 14% runtime to 12.5%
getTileInfo also reduced from 27% to 23%
* makeFastFace should use a const ref too
this trigger a const pointer need in the underlying function
Also fix some code style and prevent calculating 4 times the same position at a point
* Reduce a comparison cost for lights in updateFastFaceRow
Cache size is configurable by the meshgen_block_cache_size (default 20 MB).
New profiler stats:
- MeshUpdateQueue MapBlock cache hit %
- MeshUpdateQueue MapBlock cache size kB
Removes one type of stutter that was seen on the client when received MapBlocks
were being handled. (the "MeshMakeData::fill" stutter)
Kind of related to at least #5239
Originally preceded by these commits, now includes them:
- Move the mesh generator thread into src/mesh_generator_thread.{cpp,h}
- mesh_generator_thread.cpp: Update code style
- MeshUpdateThread: Modify interface to house a different implementation: Actual functionality will be changed by next commits.
- MeshMakeData: Add fillBlockData() interface (so that caller can fill in stuff from eg. a MapBlock cache)
* Some performance optimizations
This is globally removing some memory useless copy
* use a const ref return on std::string Settings::get to prevent data copy on getters which doesn't need to copy it
* pass some stack created strings to static const as they are not modified anywhere
* Camera: return nametags per const ref instead of a list pointer, we only need to read it
* INodeDefManager: getAll should be a result ref writer instead of a return copy
* INodeDefManager: getAlias should return a const std::string ref
* Minimap: unroll a Scolor creation in blitMinimapPixersToImageRadar to prvent many variable construct/destruct which are unneeded (we rewrite the content in the loop)
* CNodeDefManager::updateAliases: prevent a idef getall copy
* Profiler: constness
* rollback_interface: create real_name later, and use const ref
* MapBlockMesh updateFastFaceRow: unroll TileSpec next_tile, which has a cost of 1.8% CPU due to variable allocation/destruction,
* MapBlockMesh updateFastFaceRow: copy next_tile to tile only if it's a different tilespec
* MapBlockMesh updateFastFaceRow: use memcpy to copy next_lights to lights to do it in a single cpu operation
- Increase ContentFeatures serialization version
- Color property and palettes for nodes
- paramtype2 = "color", "colored facedir" or "colored wallmounted"
* Environment code refactoring
* Cleanup includes & class declarations in client & server environment to improve build speed
* ServerEnvironment::m_gamedef is now a pointer to Server instead of IGameDef, permitting to cleanup many casts.
* Cleanup IGameDef
* Move ITextureSource* IGameDef::getTextureSource() to Client only.
* Also move ITextureSource *IGameDef::tsrc() helper
* drop getShaderSource, getSceneManager, getSoundManager & getCamera abstract call
* drop unused emerge() call
* cleanup server unused functions (mentionned before)
* Drop one unused parameter from ContentFeatures::updateTextures
* move checkLocalPrivilege to Client
* Remove some unnecessary casts
* create_formspec_menu: remove IWritableTextureSource pointer, as client already knows it
* Fix some comments
* Change required IGameDef to Server/Client pointers
* Previous change that game.cpp sometimes calls functions with Client + InventoryManager + IGameDef in same functions but it's the same objects
* Remove duplicate Client pointer in GUIFormSpecMenu::GUIFormSpecMenu
* drop ClientMap::sectorWasDrawn which is unused
Move debug streams to log.cpp|h
Move GUI-related globals to clientlauncher
Move g_settings and g_settings_path to settings.cpp|h
Move g_menuclouds to clouds.cpp|h
Move g_profiler to profiler.cpp|h
Directely or indirectly optimises the following functions:
* MapBlockMesh::MapBlockMesh
* MapBlockMesh::getTileInfo
* MapBlockMesh::makeFastFace
* MapBlockMesh::getSmoothLightCombined
- Don't create and cache an extruded mesh for every (non-node) item.
Instead use a single one per image resolution.
- For cubic nodes reuse a single wield mesh too
- Improve lighting of the wielded item
- Increase far value of wield mesh scene camera, fixes#1770
- Also includes some minor refactorings of Camera and GenericCAO.
