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
|
#include <cuda_runtime.h>
#include <math.h>
#include <stdio.h>
#ifndef M_PI
#define M_PI 3.14159265358979323846
#endif
__global__ void torusVertexKernel(float* vertices, int numVertices, float time, float majorRadius, float minorRadius) {
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if (idx < numVertices) {
int vertexOffset = idx * 6;
int segments = (int)sqrt((float)numVertices);
int rings = segments;
int u = idx % segments;
int v = idx / segments;
if (v >= rings) return;
float theta = 2.0f * M_PI * u / segments;
float phi = 2.0f * M_PI * v / rings;
theta += time * 0.5f;
phi += time * 0.3f;
float cosTheta = cosf(theta);
float sinTheta = sinf(theta);
float cosPhi = cosf(phi);
float sinPhi = sinf(phi);
float x = (majorRadius + minorRadius * cosPhi) * cosTheta;
float y = (majorRadius + minorRadius * cosPhi) * sinTheta;
float z = minorRadius * sinPhi;
vertices[vertexOffset + 0] = x;
vertices[vertexOffset + 1] = y;
vertices[vertexOffset + 2] = z;
}
}
__global__ void torusColorKernel(float* colors, int numVertices, float time) {
int idx = blockIdx.x * blockDim.x + threadIdx.x;
if (idx < numVertices) {
int colorOffset = idx * 3;
float phase = time * 2.0f + idx * 0.1f;
colors[colorOffset + 0] = 0.5f + 0.5f * sinf(phase);
colors[colorOffset + 1] = 0.5f + 0.5f * sinf(phase + 2.1f);
colors[colorOffset + 2] = 0.5f + 0.5f * sinf(phase + 4.2f);
}
}
static float* d_vertices = nullptr;
static float* d_colors = nullptr;
static int maxVertices = 0;
static bool cudaInitialized = false;
extern "C" {
__declspec(dllexport) int initCUDA() {
if (cudaInitialized) return 1;
cudaError_t error = cudaSetDevice(0);
if (error != cudaSuccess) {
printf("CUDA Error: %s\n", cudaGetErrorString(error));
return 0;
}
// printf("CUDA Device initialized successfully\n");
cudaInitialized = true;
return 1;
}
__declspec(dllexport) void cleanupCUDA() {
if (!cudaInitialized) return;
if (d_vertices) { cudaFree(d_vertices); d_vertices = nullptr; }
if (d_colors) { cudaFree(d_colors); d_colors = nullptr; }
maxVertices = 0;
cudaInitialized = false;
cudaDeviceReset();
printf("CUDA cleaned up\n");
}
int ensureDeviceMemory(int numVertices) {
if (!cudaInitialized) return 0;
if (numVertices > maxVertices) {
if (d_vertices) cudaFree(d_vertices);
if (d_colors) cudaFree(d_colors);
size_t vertexSize = numVertices * 6 * sizeof(float);
size_t colorSize = numVertices * 3 * sizeof(float);
if (cudaMalloc(&d_vertices, vertexSize) != cudaSuccess) return 0;
if (cudaMalloc(&d_colors, colorSize) != cudaSuccess) return 0;
maxVertices = numVertices;
}
return 1;
}
__declspec(dllexport) void updateTorusVerticesCUDA(float* vertices, int numVertices, float time, float majorRadius, float minorRadius) {
if (!cudaInitialized || !ensureDeviceMemory(numVertices)) {
printf("CUDA not initialized or memory allocation failed\n");
return;
}
size_t vertexSize = numVertices * 6 * sizeof(float);
if (cudaMemcpy(d_vertices, vertices, vertexSize, cudaMemcpyHostToDevice) != cudaSuccess) {
printf("Failed to copy vertices to device\n");
return;
}
int blockSize = 256;
int gridSize = (numVertices + blockSize - 1) / blockSize;
torusVertexKernel<<<gridSize, blockSize>>>(d_vertices, numVertices, time, majorRadius, minorRadius);
if (cudaDeviceSynchronize() != cudaSuccess) {
printf("CUDA kernel execution failed\n");
return;
}
if (cudaMemcpy(vertices, d_vertices, vertexSize, cudaMemcpyDeviceToHost) != cudaSuccess) {
printf("Failed to copy vertices from device\n");
return; }
// printf("CUDA vertex update completed for %d vertices\n", numVertices);
}
__declspec(dllexport) void updateTorusColorsCUDA(float* colors, int numVertices, float time) {
if (!cudaInitialized || !ensureDeviceMemory(numVertices)) {
printf("CUDA not initialized or memory allocation failed\n");
return;
}
size_t colorSize = numVertices * 3 * sizeof(float);
if (cudaMemcpy(d_colors, colors, colorSize, cudaMemcpyHostToDevice) != cudaSuccess) {
printf("Failed to copy colors to device\n");
return;
}
int blockSize = 256;
int gridSize = (numVertices + blockSize - 1) / blockSize;
torusColorKernel<<<gridSize, blockSize>>>(d_colors, numVertices, time);
if (cudaDeviceSynchronize() != cudaSuccess) {
printf("CUDA color kernel execution failed\n");
return;
}
if (cudaMemcpy(colors, d_colors, colorSize, cudaMemcpyDeviceToHost) != cudaSuccess) {
printf("Failed to copy colors from device\n");
return; }
// printf("CUDA color update completed for %d vertices\n", numVertices);
}
} // extern "C"
|
