Subversion Repository Public Repository

Divide-Framework

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

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
/*
Copyright (c) 2014 DIVIDE-Studio
Copyright (c) 2009 Ionut Cava

This file is part of DIVIDE Framework.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software
and associated documentation files (the "Software"), to deal in the Software without restriction,
including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so,
subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

*/

#ifndef _SHADER_BUFFER_H_
#define _SHADER_BUFFER_H_

#include "Hardware/Video/Headers/RenderAPIWrapper.h"
#include "Utility/Headers/GUIDWrapper.h"
#include <boost/noncopyable.hpp>

class ShaderProgram;
class ShaderBuffer : private boost::noncopyable, public GUIDWrapper {
public:
    ShaderBuffer(bool unbound, bool persistentMapped) : GUIDWrapper(), _unbound(unbound), _primitiveSize(0),
                                                        _primitiveCount(0), _bufferSize(0), _persistentMapped(persistentMapped)
    {
    }

    virtual ~ShaderBuffer()
    {
    }

    ///Create a new buffer to hold our shader data
    virtual void Create(U32 primitiveCount, ptrdiff_t primitiveSize) {
        _primitiveCount = primitiveCount;
        _primitiveSize = primitiveSize;
        _bufferSize = primitiveSize * primitiveCount;
    }

    virtual void UpdateData(ptrdiff_t offset, ptrdiff_t size, const void *data, const bool invalidateBuffer = false) const = 0;
    inline  void SetData(const void *data) { UpdateData(0, _bufferSize, data, true); }

    virtual bool BindRange(Divide::ShaderBufferLocation bindIndex, U32 offsetElementCount, U32 rangeElementCount) const = 0;
    virtual bool Bind(Divide::ShaderBufferLocation bindIndex) const = 0;
    virtual void PrintInfo(const ShaderProgram* shaderProgram, Divide::ShaderBufferLocation bindIndex) = 0;

    inline size_t getPrimitiveSize()  const { return _primitiveSize; }
    inline U32    getPrimitiveCount() const { return _primitiveCount; }

protected:
    bool   _unbound;
    bool   _persistentMapped;
    size_t _bufferSize;
    size_t _primitiveSize;
    U32    _primitiveCount;
};

#endif

Commits for Divide-Framework/trunk/Source Code/Hardware/Video/Buffers/ShaderBuffer/Headers/ShaderBuffer.h

Diff revisions: vs.
Revision Author Commited Message
286 Diff Diff IonutCava picture IonutCava Fri 18 Jul, 2014 16:32:03 +0000

[Ionut] [[BR]]

    • very buggy commit ** [[BR]]

- New rendering pipeline: [[BR]]
— Bumped OpenGL version to 4.4 [[BR]]
— Command based rendering. GFXDevice dispatches rendering commands to their respective buffers [[BR]]
— State changing is exclusively controlled by the GFXDevice class [[BR]]
— Material properties and transform matrices are stored in shader storage buffers that are updated after culling and uploaded once per RenderPass. Objects use drawIDs to index data in shaders [[BR]]
— Removed all custom matrices and other, per shader, bind data [[BR]]
— Removed per node light data. Light culling and shader buffers will replace the old light management system [[BR]]
— Framebuffers now use the Texture class for attachments [[BR]]
— Sampler objects now work like RenderStateBlocks. GL_API handles them and stores them in a hash indexed map [[BR]]
— Removed “prepareMaterial” calls. Pre-draw material setup is now only limited to texture binding [[BR]]
— Moved immediate mode emulation and primitive rendering from API level to GFXDevice level [[BR]]
— Framebuffer rendering must be completed with a call to end() to ensure proper MSAA resolve (asserts, otherwise) [[BR]]
- Fixed hash code generation for RenderStateBlock (and now, glSamplerObject) using boost::hash_combine [[BR]]
- Temporarily disabled XML caching for materials [[BR]]

285 Diff Diff IonutCava picture IonutCava Wed 02 Jul, 2014 18:33:25 +0000

[Ionut][[BR]]
- Per SGN shader data implementation part 1.75 / 2 [[BR]]
— Upload node data into a SSBO in one single call [[BR]]
- Re-factored LightManager to upload light info and shadow info in single upload calls [[BR]]
- Improved GUIConsole load performance by using a boost::circular_buffer for initial messages [[BR]]

274 Diff Diff IonutCava picture IonutCava Fri 13 Jun, 2014 12:56:03 +0000

[Ionut] [[BR]]
- Simplified buffer based rendering (VertexBuffer & GenericVertexData are now exclusively command based) [[BR]]
- Update bone buffer to persistent mapped-SSBO system with double-buffering [[BR]]
- Simplified sky rendering and (finally) reduced the sun’s radius [[BR]]

273 Diff Diff IonutCava picture IonutCava Wed 11 Jun, 2014 20:34:00 +0000

[Ionut] [[BR]]
- Improved batching part2: [[BR]]
— Moved all bone transformation matrices upload from uniform functions in SceneNode to Shader Storage Buffers in AnimationController class with updates controlled by SceneGraphNode’s AnimationComponent [[BR]]
-— No double-buffering or efficient mapping of data yet. Just raw buffer updates [[BR]]
— Replaced all RenderStateBlock* pointers with I64 hash values [[BR]]
-— Will be used later in draw commands as per-shader state hash [[BR]]
— Moved getShaderProgram from Material to Material::ShaderInfo [[BR]]
-— Improved redundant uniform checks [[BR]]
— Improved texture unit changing in case of redundant texture bind case (i.e. skip changing the texture unit) [[BR]]
[[BR]]
- Improved float and double comparison functions using algorithms described in http://randomascii.wordpress.com/2012/02/13/dont-store-that-in-a-float/ [[BR]]
- Added a new utility function, DIVIDE_ASSERT, to more easily bind an assert check with a specific message [[BR]]
- Added missing files from Forward+ renderer’s initial code [[BR]]

272 Diff Diff IonutCava picture IonutCava Wed 11 Jun, 2014 11:30:51 +0000

[Ionut] [[BR]]
- Added initial code for Forward+ rendering (grid based light culling) [[BR]]
- Simplified Light class [[BR]]
- Moved more light uniform data to ShaderBuffers [[BR]]
- Double buffered grass rendering [[BR]]
- Added a better attribute interface to GenericVertexData [[BR]]
- Fixed screenshot capture [[BR]]
- Fixed texture double-bind checks [[BR]]

268 IonutCava picture IonutCava Wed 04 Jun, 2014 10:20:27 +0000

[IonutCava] [[BR]]
- Terrain and vegetation updates: [[BR]]
— Use standard phong lighting model [[BR]]
— Separate grass patches per chunk [[BR]]
— Reduce number of required texture bind points for terrain by adding both textures and normal maps in a single texture array [[BR]]
— Draw terrain with 2 drawcalls (using glMultiDrawElements twice, once for LoD == 0 and once for LoD > 0) [[BR]]
[[BR]]
- Buffer management updates: [[BR]]
— Unified buffer binding system [[BR]]
— Made Uniform buffer objects visible to all classes and moved lightUBO to LightManager, matricesUBO to GFXDevice etc [[BR]]
— Added initial support for Shader Buffer Objects [[BR]]
— Moved viewport management to GFXDevice class [[BR]]
[[BR]]
- Rendering improvements: [[BR]]
— Added Hi-Z occlusion culling to instanceCull shader (using Hi-Z depth buffer constructed in GFXDevice class) [[BR]]
— Resolved issues with shadow maps overriding texture bind locations [[[BR]]
— Improved shader and shader program loading [[BR]]
— Improved material transparency support [[BR]]
— Added buffer lock manager class from “Approaching Zero Driver Overhead in OpenGL (Presented by NVIDIA)” speech at GDC14 [[BR]]