forked from chrisyeshi/VolRen
-
Notifications
You must be signed in to change notification settings - Fork 0
/
volrenraycastcuda.cpp
172 lines (151 loc) · 6.73 KB
/
volrenraycastcuda.cpp
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
#include "volrenraycastcuda.h"
#include "volrenraycastcuda.cuda.h"
#include <cassert>
#include "imagepbo.h"
#include "volumegl.h"
//
//
// cuda error checking
//
//
#define cc(ans) { gpuAssert((ans), __FILE__, __LINE__); }
static void gpuAssert(cudaError_t code, const char *file, int line, bool abort=true)
{
if (code != cudaSuccess)
{
fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line);
if (abort) exit(code);
}
}
//
//
// VolRenRaycastCuda
//
//
namespace yy {
namespace volren {
VolRenRaycastCuda::VolRenRaycastCuda()
: TFIntegrated<VolRenRaycast>(Method_Raycast_CUDA)
, outPBO(0)
, entryRes(NULL), exitRes(NULL), outRes(NULL)
, texWidth(defaultFBOSize), texHeight(defaultFBOSize)
, volRes(0)
, tfFullRes(0), tfBackRes(0)
{
}
VolRenRaycastCuda::~VolRenRaycastCuda()
{
if (entryRes) cc(cudaGraphicsUnregisterResource(entryRes));
if (exitRes) cc(cudaGraphicsUnregisterResource(exitRes));
if (outRes) cc(cudaGraphicsUnregisterResource(outRes));
if (volRes) cc(cudaGraphicsUnregisterResource(volRes));
if (tfFullRes) cc(cudaGraphicsUnregisterResource(tfFullRes));
if (tfBackRes) cc(cudaGraphicsUnregisterResource(tfBackRes));
QOpenGLFunctions f(QOpenGLContext::currentContext());
if (0 != outPBO) f.glDeleteBuffers(1, &outPBO);
}
void VolRenRaycastCuda::initializeGL()
{
VolRenRaycast::initializeGL();
updateCUDAResources();
}
void VolRenRaycastCuda::resize(int w, int h)
{
VolRenRaycast::resize(w, h);
updateCUDAResources();
}
void VolRenRaycastCuda::setTF(const mslib::TF &tf, bool preinteg, float stepsize, Filter filter)
{
TFIntegrated<VolRenRaycast>::setTF(tf, preinteg, stepsize, filter);
if (tfFullRes) cc(cudaGraphicsUnregisterResource(tfFullRes));
cc(cudaGraphicsGLRegisterImage(&tfFullRes, tfInteg->getTexFull()->textureId(), GL_TEXTURE_2D, cudaGraphicsRegisterFlagsReadOnly));
if (tfBackRes) cc(cudaGraphicsUnregisterResource(tfBackRes));
cc(cudaGraphicsGLRegisterImage(&tfBackRes, tfInteg->getTexBack()->textureId(), GL_TEXTURE_2D, cudaGraphicsRegisterFlagsReadOnly));
}
std::shared_ptr<ImageAbstract> VolRenRaycastCuda::output() const
{
return std::make_shared<ImagePBO>(outPBO, frustum.getTextureWidth(), frustum.getTextureHeight());
}
void VolRenRaycastCuda::raycast(const QMatrix4x4&, const QMatrix4x4& matView, const QMatrix4x4&)
{
cc(cudaGraphicsMapResources(1, &entryRes, 0));
cc(cudaGraphicsMapResources(1, &exitRes, 0));
cc(cudaGraphicsMapResources(1, &outRes, 0));
cc(cudaGraphicsMapResources(1, &volRes, 0));
cc(cudaGraphicsMapResources(1, &tfFullRes, 0));
cc(cudaGraphicsMapResources(1, &tfBackRes, 0));
cudaArray *entryArr, *exitArr, *volArr, *tfFullArr, *tfBackArr;
float *outPtr;
size_t nBytes;
cc(cudaGraphicsSubResourceGetMappedArray(&entryArr, entryRes, 0, 0));
cc(cudaGraphicsSubResourceGetMappedArray(&exitArr, exitRes, 0, 0));
cc(cudaGraphicsResourceGetMappedPointer((void**)&outPtr, &nBytes, outRes));
cc(cudaGraphicsSubResourceGetMappedArray(&volArr, volRes, 0, 0));
cc(cudaGraphicsSubResourceGetMappedArray(&tfFullArr, tfFullRes, 0, 0));
cc(cudaGraphicsSubResourceGetMappedArray(&tfBackArr, tfBackRes, 0, 0));
static std::map<Filter, cudaTextureFilterMode> vr2cu
= { { Filter_Linear, cudaFilterModeLinear }
, { Filter_Nearest, cudaFilterModePoint } };
assert(vr2cu.count(tfFilter) > 0);
updateCUDALights(matView);
cudacast(vol()->w(), vol()->h(), vol()->d(), volArr,
tfInteg->getTexFull()->width(), tfInteg->getTexFull()->height(), stepsize, vr2cu[tfFilter], tfFullArr, tfBackArr,
scalarMin, scalarMax,
frustum.getTextureWidth(), frustum.getTextureHeight(), entryArr, exitArr, outPtr);
cc(cudaGraphicsUnmapResources(1, &tfBackRes, 0));
cc(cudaGraphicsUnmapResources(1, &tfFullRes, 0));
cc(cudaGraphicsUnmapResources(1, &volRes, 0));
cc(cudaGraphicsUnmapResources(1, &entryRes, 0));
cc(cudaGraphicsUnmapResources(1, &exitRes, 0));
cc(cudaGraphicsUnmapResources(1, &outRes, 0));
}
void VolRenRaycastCuda::volumeChanged()
{
if (volRes) cc(cudaGraphicsUnregisterResource(volRes));
cc(cudaGraphicsGLRegisterImage(&volRes, volume->getTexture()->textureId(), GL_TEXTURE_3D, cudaGraphicsRegisterFlagsReadOnly));
}
void VolRenRaycastCuda::updateCUDAResources()
{
QOpenGLFunctions f(QOpenGLContext::currentContext());
// update entry and exit resources
if (entryRes) cc(cudaGraphicsUnregisterResource(entryRes));
cc(cudaGraphicsGLRegisterImage(&entryRes, *frustum.entryTexture(), GL_TEXTURE_2D, cudaGraphicsRegisterFlagsReadOnly));
if (exitRes) cc(cudaGraphicsUnregisterResource(exitRes));
cc(cudaGraphicsGLRegisterImage(&exitRes, *frustum.exitTexture(), GL_TEXTURE_2D, cudaGraphicsRegisterFlagsReadOnly));
// update output PBO and CUDA resource
// clean up
if (0 != outPBO) f.glDeleteBuffers(1, &outPBO);
// pbo
f.glGenBuffers(1, &outPBO);
f.glBindBuffer(GL_PIXEL_UNPACK_BUFFER, outPBO);
f.glBufferData(GL_PIXEL_UNPACK_BUFFER, 3 * frustum.getTextureWidth() * frustum.getTextureHeight() * sizeof(float), NULL, GL_STREAM_COPY);
f.glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0);
// register cuda resource
if (outRes) cc(cudaGraphicsUnregisterResource(outRes));
cc(cudaGraphicsGLRegisterBuffer(&outRes, outPBO, cudaGraphicsMapFlagsWriteDiscard));
}
void VolRenRaycastCuda::updateCUDALights(const QMatrix4x4& matView)
{
CudaLight cudaLights[10];
for (unsigned int i = 0; i < lights.size(); ++i)
{
QVector3D lightDir(matView.inverted() * QVector4D(lights[i].direction, 0.f));
lightDir.normalize();
cudaLights[i].direction.x = lightDir.x();
cudaLights[i].direction.y = lightDir.y();
cudaLights[i].direction.z = lightDir.z();
cudaLights[i].ambient.x = lights[i].color.x() * lights[i].ambient;
cudaLights[i].ambient.y = lights[i].color.y() * lights[i].ambient;
cudaLights[i].ambient.z = lights[i].color.z() * lights[i].ambient;
cudaLights[i].diffuse.x = lights[i].color.x() * lights[i].diffuse;
cudaLights[i].diffuse.y = lights[i].color.y() * lights[i].diffuse;
cudaLights[i].diffuse.z = lights[i].color.z() * lights[i].diffuse;
cudaLights[i].specular.x = lights[i].color.x() * lights[i].specular;
cudaLights[i].specular.y = lights[i].color.y() * lights[i].specular;
cudaLights[i].specular.z = lights[i].color.z() * lights[i].specular;
cudaLights[i].shininess = lights[i].shininess;
}
cudaSetLights(lights.size(), cudaLights);
}
} // namespace volren
} // namespace yy