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
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
|
inline void checkErr(cl_int err, const char * name) {
if (err != CL_SUCCESS) {
std::cerr << "ERROR: " << name << " (" << err << ")" << std::endl;
exit(EXIT_FAILURE);
}
}
inline void checkErr(cl_program prog, void* name) {
std::cerr << "ERROR: " << name << std::endl;
}
CPURenderComponent::CPURenderComponent(math::Vec3f position,math::Vec3f size, math::Vec3f origin, bool useThread, RenderWindow& rw)
: Component(position,size, origin, useThread), window(rw), view (rw.getView()) {
redBuffer.resize(size.x * size.y);
blueBuffer.resize(size.x * size.y);
greenBuffer.resize(size.x * size.y);
alphaBuffer.resize(size.x * size.y);
depthBuffer.resize(size.x * size.y);
this->size.x = rw.getSize().x;
this->size.y = rw.getSize().y;
view = window.getView();
fbTile = new Tile(&fbTexture, position, size, sf::IntRect(0, 0, size.x, size.y));
fbTile->setCenter(view.getPosition());
cl_int err;
std::vector< cl::Platform > platformList;
cl::Platform::get(&platformList);
checkErr(platformList.size()!=0 ? CL_SUCCESS : -1, "cl::Platform::get");
std::cerr << "Platform number is: " << platformList.size() << std::endl;std::string platformVendor;
platformList[0].getInfo((cl_platform_info)CL_PLATFORM_VENDOR, &platformVendor);
std::cerr << "Platform is by: " << platformVendor << "\n";
cl_context_properties cprops[3] = {CL_CONTEXT_PLATFORM, (cl_context_properties)(platformList[0])(), 0};
clcontext = cl::Context (
CL_DEVICE_TYPE_ALL,
cprops,
NULL,
NULL,
&err);
devices = clcontext.getInfo<CL_CONTEXT_DEVICES>();
checkErr(devices.size() > 0 ? CL_SUCCESS : -1, "devices.size() > 0");
std::string prog =
"//Multiply two 4*4 matrices."
"float16 multMat (float16 matA, float16 matB) {"
"float16 matC;"
"for (int x = 0; x < 4; ++x) {"
"for (int y = 0; y < 4; ++y) {"
"float value = 0;"
"for (int k = 0; k < 4; ++k) {"
"float elementA = matA[y * 4 + k];"
"float elementB = matB[k * 4 + x];"
"value += elementA * elementB;"
"}"
"matC[y * 4 + x] = value;"
"}"
"}"
"return matC;"
"}"
"//Multiply a 4*4 matrix with a vec4."
"float4 multVec (float16 matA, float4 vecB) {"
"float4 vecC;"
"for (int i = 0; i < 4; ++i) {"
"float value = 0;"
"for (int j = 0; j < 4; ++j) {"
"value += vecB[j] * matA[i][j];"
"}"
"vecC[i] = value;"
"}"
"return vecC;"
"}"
"//Transpose a 4*4 matrix."
"float16 transpose(float16 matA) {"
"float16 matT"
"for (int i = 0; i < 4; ++i) {"
"for (int j = 0; j < 4; ++j) {"
"matT[i][j] = matA[j][i];"
"}"
"}"
"return matT;"
"}"
"float det (float4 mat) {"
"return mat[0] * mat[2] - mat[1] * mat[3];"
"}"
"float min (float3 vec) {"
"float cmin = vec.x;"
"for (int i = 1; i < 3; ++i) {"
"if (vec[i] < cmin)"
"cmin = vec[i];"
"}"
"return cmin;"
"}"
"float max (float3 vec) {"
"float cmax = vec.x;"
"for (int i = 1; i < 3; ++i) {"
"if (vec[i] > cmax)"
"cmax = vec[i];"
"}"
"return cmax;"
"}"
"//Convert vertex position from object space toviewport space."
"__kernel void vertexShader(__global float* vPosX, __global float* vPosY, __global float* vPosZ, __global float* vPosW,"
"__global unsigned int* indices, __global unsigned int numIndices, __global unsigned int* baseIndices,"
"__global unsigned int* baseVertices, __global unsigned int* nbVerticesPerFaces,"
"__global float* transfMatrices, __global float16 projMatrix, __global float16 viewMatrix,"
"__global float16 viewportMatrix) {"
"size_t tid = get_global_id(0);"
"int instanceID = tid / nbVerticesPerFace[0];"
"int offset = tid % nbVerticesPerFace[0];"
"float16 transfMatrix;"
"float4 position = (float4) (vPosX[baseVertices[instanceID]+indices[baseIndices[instanceID]+offset], vPosY[baseVertices[instanceID]+indices[baseIndices[instanceID]+offset],"
"vPosZ[baseVertices[instanceID]+indices[baseIndices[instanceID]+offset], vPosW[baseVertices[instanceID]+indices[baseIndices[instanceID]+offset]);"
"for (int i = 0; i < 16; i++) {"
"transfMatrix[i] = transfMatrices[instanceID*16+i];"
"}"
"float4 worldcoords = multVec(transfMatrix, position);"
"float4 viewcoords = multVec(viewMatrix, worldcoords);"
"float4 clipcoords = multVec(projMatrix, viewcoords);"
"float4 ndcCoords = clipcoords / clipcoords.w;"
"position = multVec(viewportMatrix, ndcCoords);"
"vPosX[tid] = abs(position.x);"
"vPosY[tid] = abs(position.y);"
"vPosZ[tid] = abs(position.z);"
"vPosW[tid] = position.w;"
"}";
cl::Program::Sources source(
1,
std::make_pair(prog.c_str(), prog.length()+1));
cl::Program program(clcontext, source);
err = program.build(devices,"",&checkErr);
checkErr(err, "Program::build()");
if (err != CL_SUCCESS) {
cl_int status, logSize;
std::string programLog;
program.getBuildInfo(devices[0],CL_PROGRAM_BUILD_LOG, &programLog);
std::cerr<<programLog<<std::endl;
}
clkvertexShader = cl::Kernel(program, "vertexShader", &err);
checkErr(err, "Kernel::Kernel()");
checkErr(err, "Kernel::setArg()");
View& view = window.getView();
ViewportMatrix vpm;
vpm.setViewport(math::Vec3f(view.getViewport().getPosition().x, view.getViewport().getPosition().y, 0),
math::Vec3f(view.getViewport().getWidth(), view.getViewport().getHeight(), 1));
std::array<float, 16> projMatrix = view.getProjMatrix().getMatrix().toGlMatrix();
std::array<float, 16> viewMatrix = view.getViewMatrix().getMatrix().toGlMatrix();
std::array<float, 16> viewportMatrix = vpm.getMatrix().toGlMatrix();
cprojMatrixBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, projMatrix.size(), projMatrix.data(), &err);
checkErr(err, "Buffer::Buffer()");
cviewMatrixBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, viewMatrix.size(), viewMatrix.data(), &err);
checkErr(err, "Buffer::Buffer()");
cvpMatrixBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, viewportMatrix.size(), viewportMatrix.data(), &err);
checkErr(err, "Buffer::Buffer()");
err = clkvertexShader.setArg(10, cprojMatrixBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(11, cviewMatrixBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(12, cvpMatrixBuffer);
checkErr(err, "Kernel::setArg()");
}
void CPURenderComponent::drawNextFrame() {
for (unsigned int i = 0; i < m_instances.size(); i++) {
std::vector<float> transformMatrices;
for (unsigned int j = 0; j < m_instances[i]->getTransforms().size(); j++) {
float* tmatrix = m_instances[i]->getTransforms()[j].get().getGlMatrix();
for (unsigned int n = 0; n < 16; n++) {
transformMatrices.push_back(tmatrix[n]);
}
}
cl_int err;
ctransfMatrixBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, transformMatrices.size(), transformMatrices.data(), &err);
checkErr(err, "Buffer::Buffer()");
math::Matrix4f texM = (m_instances[i]->getMaterial().getTexture() != nullptr) ?
m_instances[i]->getMaterial().getTexture()->getTextureMatrix() :
math::Matrix4f();
std::vector<float> vPosX = m_instances[i]->getVertexArray().m_vPosX;
std::vector<float> vPosY = m_instances[i]->getVertexArray().m_vPosY;
std::vector<float> vPosZ = m_instances[i]->getVertexArray().m_vPosZ;
std::vector<float> vPosW = m_instances[i]->getVertexArray().m_vPosW;
std::vector<unsigned char> vcRed = m_instances[i]->getVertexArray().m_vcRed;
std::vector<unsigned char> vcBlue = m_instances[i]->getVertexArray().m_vcBlue;
std::vector<unsigned char> vcGreen = m_instances[i]->getVertexArray().m_vcGreen;
std::vector<unsigned char> vcAlpha = m_instances[i]->getVertexArray().m_vcAlpha;
std::vector<unsigned int> ctx = m_instances[i]->getVertexArray().m_ctX;
std::vector<unsigned int> cty = m_instances[i]->getVertexArray().m_ctY;
std::vector<unsigned int> indexes = m_instances[i]->getVertexArray().m_indexes;
std::vector<unsigned int> numIndexes = m_instances[i]->getVertexArray().m_numIndexes;
std::vector<unsigned int> baseIndexes = m_instances[i]->getVertexArray().m_baseIndexes;
std::vector<unsigned int> baseVertices = m_instances[i]->getVertexArray().m_baseVertices;
std::vector<unsigned int> nbVerticesPerFaces;
std::array<float, 16> texMatrix = texM.toGlMatrix();
nbVerticesPerFaces.push_back(m_instances[i]->getVertexArray().nbVerticesPerFace);
nbVertices.push_back(m_instances[i]->getVertexArray().getVertexCount());
cvposXBuffer = cl::Buffer(clcontext, CL_MEM_READ_WRITE | CL_MEM_USE_HOST_PTR, vPosX.size(), vPosX.data(), &err);
checkErr(err, "Buffer::Buffer()");
cvposYBuffer = cl::Buffer(clcontext, CL_MEM_READ_WRITE | CL_MEM_USE_HOST_PTR, vPosX.size(), vPosY.data(), &err);
checkErr(err, "Buffer::Buffer()");
cvposZBuffer = cl::Buffer(clcontext, CL_MEM_READ_WRITE | CL_MEM_USE_HOST_PTR, vPosX.size(), vPosZ.data(), &err);
checkErr(err, "Buffer::Buffer()");
cvposWBuffer = cl::Buffer(clcontext, CL_MEM_READ_WRITE | CL_MEM_USE_HOST_PTR, vPosX.size(), vPosW.data(), &err);
checkErr(err, "Buffer::Buffer()");
cvindexesBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, indexes.size(), indexes.data(), &err);
checkErr(err, "Buffer::Buffer()");
cnumIndexesBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, numIndexes.size(), numIndexes.data(), &err);
checkErr(err, "Buffer::Buffer()");
cbaseIndexesBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, baseIndexes.size(), baseIndexes.data(), &err);
checkErr(err, "Buffer::Buffer()");
cbaseVerticesBuffer = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, baseVertices.size(), baseVertices.data(), &err);
checkErr(err, "Buffer::Buffer()");
cNbVerticesPerFaces = cl::Buffer(clcontext, CL_MEM_READ_ONLY | CL_MEM_USE_HOST_PTR, nbVerticesPerFaces.size(), nbVerticesPerFaces.data(), &err);
checkErr(err, "Buffer::Buffer()");
err = clkvertexShader.setArg(0, cvposXBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(1, cvposYBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(2, cvposZBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(3, cvposWBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(4, cvindexesBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(5, cnumIndexesBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(6, cbaseIndexesBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(7, cbaseVerticesBuffer);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(8, cNbVerticesPerFaces);
checkErr(err, "Kernel::setArg()");
err = clkvertexShader.setArg(9, ctransfMatrixBuffer);
checkErr(err, "Kernel::setArg()");
cl::CommandQueue clqueue (clcontext, devices[0], 0, &err);
checkErr(err, "CommandQueue::CommandQueue()");
cl::Event event;
cl_int workgroupSize = m_instances[i]->getVertexArray().getVertexCount();
err = clqueue.enqueueNDRangeKernel(clkvertexShader,
1, workgroupSize,
1, NULL, &event);
checkErr(err, "ComamndQueue::enqueueNDRangeKernel()");
event.wait();
err = clqueue.enqueueReadBuffer(cvposXBuffer,CL_TRUE,0,workgroupSize,vPosX.data());
checkErr(err, "ComamndQueue::enqueueReadBuffer()");
err = clqueue.enqueueReadBuffer(cvposYBuffer,CL_TRUE,0,workgroupSize,vPosY.data());
checkErr(err, "ComamndQueue::enqueueReadBuffer()");
err = clqueue.enqueueReadBuffer(cvposZBuffer,CL_TRUE,0,workgroupSize,vPosZ.data());
checkErr(err, "ComamndQueue::enqueueReadBuffer()");
err = clqueue.enqueueReadBuffer(cvposWBuffer,CL_TRUE,0,workgroupSize,vPosW.data());
checkErr(err, "ComamndQueue::enqueueReadBuffer()");
for (unsigned int i = 0; i < workgroupSize; i++)
std::cout<<vPosX[i]<<" "<<vPosY[i]<<" "<<vPosZ[i]<<" "<<vPosW[i]<<std::endl;
}
} |
Erreur lors de la compilation du kernel
Répondre avec citation
Partager