Subversion Repository Public Repository


This repository has no backups
This repository's network speed is throttled to 100KB/sec

Commits for Divide-Framework/trunk/Source Code/Rendering/RenderPass/Headers/RenderPass.h

Diff revisions: vs.
Revision Author Commited Message
1029 Diff Diff IonutCava picture IonutCava Tue 30 Jan, 2018 17:28:39 +0000

- Update copyright notices

929 Diff Diff IonutCava picture IonutCava Sat 12 Aug, 2017 17:34:44 +0000

- More work on finishing up new PushConstant / Command buffer rendering system:
— Everything compiles and runs but with graphical and performance issues

863 Diff Diff IonutCava picture IonutCava Wed 15 Feb, 2017 16:27:26 +0000

- Split Resource class into Resource and CachedResource
— CachedResource is the only resource held in a ResourceCache and loaded from a ResourceDescriptor
— CachedResource is mapped by its creation descriptor’s hash value, instead of the resource name
-— Multiple cached resources can now exist with the same name if the have different properties, for example.

852 Diff Diff IonutCava picture IonutCava Tue 07 Feb, 2017 17:17:27 +0000

- Separate Pass Type from Render Stage:
— Render Stage: Display, Shadow, Reflection, Refraction, etc
— Pass Type: Depth pass, Colour Pass, etc
- ToDo: Bug fixes for issues introduced with this change

836 Diff Diff IonutCava picture IonutCava Fri 27 Jan, 2017 14:59:56 +0000

- Update copyright notice

807 Diff Diff IonutCava picture IonutCava Mon 09 Jan, 2017 17:06:26 +0000

- Split PlatformDefines.h into 2. Move data type defines to PlatformDataTypes.h
- Continue to improve velocity map calculation
- Remove “destroy()” calls from most classes. Attempt to move as many classes to the RAII system.

804 Diff Diff IonutCava picture IonutCava Thu 01 Dec, 2016 17:20:59 +0000

- Singleton elimination update Part I: get it to compile
— The following classes are no longer Singletons: GFXDevice, GL_API, DX_API, SFXWrapper, FmodWrapper, SDLWrapper, ALWrapper, PXDevice, InputInterface, RenderPassManager, SceneManager and ResourceManager;
— Refactor system to a Context / Component based implementation (Pass relevant context to objects: e.g. GFXDevice object to Textures, GUI to GUIElements, etc)
— Make devices and managers components of the kernel
— Allow multiple Resource caches to co-exist. This may prove useful for later when a more fragmented memory model is need (per frame / per scene / global caches / etc)

  • next steps: part II – cleanup/refactor new code, part III – optimise code, part IV – remove remaining Singletons (e.g. Application, ParamHandler, FrameListenerManager, Recast, PostFX and DebugInterface)
801 Diff Diff IonutCava picture IonutCava Sun 27 Nov, 2016 21:28:01 +0000

- More performance analysis guided optimizations.
- Some refactoring to allow later removal of Singletons status for: GFXDevice, SFXDevice, PXDevice, GUI and Input.
— Pass a PlatformContext around objects to access these classes.

773 Diff Diff IonutCava picture IonutCava Tue 04 Oct, 2016 15:53:18 +0000

- Split GenericDrawCommand gathering system into Init/Update to facilitate future threading support
- Fix drawCount() == 0 case in submitRenderCommand
- Make _cameraPool (more) threadsafe

760 Diff Diff IonutCava picture IonutCava Tue 09 Aug, 2016 16:30:12 +0000

- Complete rewrite of RenderPass system:
— Allow every rendering pass (reflection, environment, etc) to use the same rendering system: occlusion culling, light culling, etc
— Define render passes by a param based token similar to modern API systems (DX12/Vulkan)
— W.I.P.: Depth rendering slightly wrong; Environment Mapping ,reflection and refraction rendering currently disabled
- Add compile time variables for DEBUG/PROFILE/RELEASE builds to validate all code for every build

734 Diff Diff IonutCava picture IonutCava Tue 21 Jun, 2016 16:28:11 +0000

- Added a render target pool:
— All render targets are now allocated (and tracked) by the GFXDevice.
— Allocation routines return a slim RenderTargetHandle object that will be used later on in draw commands.

- Added a refraction render pass that will handle translucent objects that need refracted textures instead of clear transparency
— Both reflective nodes and refractive nodes are not gathered properly at this point

- All API specific objects (buffers, textures, etc) are now initialized in the GFXDevice class (see GFXDeviceObjects.cpp)
— This allowed for the slimming of the RenderAPIWrapper interface

- Improved ENABLE_GPU_VALIDATION macro utilization:
— It enabled shadow map split plane debugging
— It bypasses loading shaders from text cache

727 Diff Diff IonutCava picture IonutCava Tue 14 Jun, 2016 16:01:38 +0000

- Remove node and cmd buffers from GFXDevice and add them as a new BufferData struct to RenderPass class.
— Each render pass holds its own buffers
- Improvements / fixes to the CSM code
- Added a global toggle to enable/disable custom memory allocators

698 Diff Diff IonutCava picture IonutCava Mon 09 May, 2016 20:21:34 +0000

- Remove all anaglyph rendering code. nVidia gained 1 sympathy point with me for VRWorks:
— Single Pass Stereo completely removes the need for double rendering the scene for each eye.
— LiquidVR also support similar technology
-— Once VR development starts for Divide, documentation for both SDKs should be pretty sorted.

648 Diff Diff IonutCava picture IonutCava Sun 21 Feb, 2016 16:32:52 +0000

- Update copyright notice
- Move BoundingBox and BoundingSphere to a new BoundsComponent
— Add a temp hack in SceneGraphNode to update these

645 Diff Diff IonutCava picture IonutCava Mon 15 Feb, 2016 16:44:16 +0000

- Added frustum planes to the GFXDevice UBO
- Occlusion culling updates

644 Diff Diff IonutCava picture IonutCava Sun 14 Feb, 2016 15:04:17 +0000

- Fix ZPrePass on AMD hardware:
— Use invariant gl_Position;
— USe early fragment test for non-alpha blended entities
— Add a toggle to disable/enable ZPrePass at compile time
- Add Occlusion culling debugging feature togglable at compile time
- Initial code for high prio materials (used to always calc reflections, compute shaders immediately, etc)

634 Diff Diff IonutCava picture IonutCava Wed 27 Jan, 2016 20:41:59 +0000

- Fix 2 issues with previous commit

629 Diff Diff IonutCava picture IonutCava Thu 21 Jan, 2016 17:08:04 +0000

- New environment mapping system:
— Use previous frame’s list of nodes sorted front to back
— Update reflection for each if they meet the necessary criteria
- Cleanup particle system classes
- Cleanup GenericVertexData class
- Improve argument forwarding to frustum culling tasks

612 Diff Diff IonutCava picture IonutCava Sun 20 Dec, 2015 17:47:26 +0000

- Z-PrePass system improvements
— Depth only rendertarget and both eye render targets (screen and anaglyph) share the same depth buffer
— Screen render targets don’t clear or write the/to the depth buffer

560 Diff Diff IonutCava picture IonutCava Wed 11 Nov, 2015 16:11:13 +0000

- Implement a proper RenderPassManager that can hold and execute, in order, multiple rendering passes, with each pass able to handle multiple rendering stages:
— Added a RenderPass for shadowmapping that gets executed first
— Added a RenderPass for environment mapping that gets executed second
— Added a RenderPass for normal rendering that gets executed third and processes 2 stages: Z_PRE_PASS and DISPLAY
— All RenderPasses get executed twice if anaglyph rendering is enableDebugRender
- Added environment map auto-generation. The second render pass creates a reflection cube map used for reflections
- Cleaned up Kernel->GFXDevice->SceneManager relation for scene rendering
- Moved Renderer to SceneManager as it is not related to the GFXDevice in any way and it’s implementation can be very scene dependent.

540 Diff Diff IonutCava picture IonutCava Wed 28 Oct, 2015 19:06:28 +0000

- Remove the “Platform” folder from the Platform project as the name is redundant. Move files one level up

523 Diff Diff IonutCava picture IonutCava Wed 07 Oct, 2015 16:11:44 +0000

- Performance optimizations:
— Change GFXDevice’s vector of RenderPackages to a standalone class that handles its own memory, thus avoiding the hit from clear()/push_back()
— Also minimizes memory allocation in each render pass
- Remove the extra WaitForLockedRange call in LockRange command. It was added to work around a bug that doesn’t happen any more due to better buffer usage.
- Add some vegetation generation notes for future updates

442 Diff Diff IonutCava picture IonutCava Wed 20 May, 2015 15:25:02 +0000

- Better platform handling:
— All platform specific code moved in 3 separate files: PlatformDefinesWindows / PlatformDefinesUnix / PlatformDefinesApple
— Added CPU and RAM check at startup to make sure we meet minimum requirements to run properly
- Smart pointers should never used MemoryManager_NEW for allocating memory (MemoryManager_NEW tracks allocations but must be matched with a DELETE call in order to no report leaks). Corrected.
- Zombification of Singletons is only enabled in Debug builds
- Fixed a bug in Task.cpp (nowTime check) that prevented some tasks to run on the first call (changed a ‘>’ check to a ‘>=’)
— This also fixed bounding box computations for skinned submeshes
- Improved dead zone handling for joysticks

353 Diff Diff IonutCava picture IonutCava Tue 17 Feb, 2015 17:08:27 +0000

- Compatibility updates for Boost 1.57 (some warnings still exist for now)
- Fixed some non-conforming code
- Added ALL OF THE MISSING FILES FROM THE LAST 15 commits ...........

350 IonutCava picture IonutCava Tue 10 Feb, 2015 16:25:39 +0000

- Simplified some Util-namespace level functions by using STL algorithms where possible
- Formatted the entire codebase using clang-format with the following style: "{BasedOnStyle: Google, IndentWidth: 4}"
- Attempted to make ProfileTimer thread-safe (temp fix until std::chrono works properly in Visual Studio)