int main(int argc, char *argv[])
{
perf_parse_args(argc, argv);
std::cout << "size: " << PERF_N << std::endl;
::cl::Device device = bolt::cl::control::getDefault().getDevice();
std::cout << "device: " << device.getInfo<CL_DEVICE_NAME>() << std::endl;
// create host vector
std::vector<int> h_vec = generate_random_vector<int>(PERF_N);
// create device vector
bolt::cl::device_vector<int> d_vec(PERF_N);
perf_timer t;
for(size_t trial = 0; trial < PERF_TRIALS; trial++){
// transfer data to the device
bolt::cl::copy(h_vec.begin(), h_vec.end(), d_vec.begin());
t.start();
bolt::cl::sort(d_vec.begin(), d_vec.end());
t.stop();
}
std::cout << "time: " << t.min_time() / 1e6 << " ms" << std::endl;
// transfer data back to host
bolt::cl::copy(d_vec.begin(), d_vec.end(), h_vec.begin());
return 0;
}
double NonRigid::computeArap(VectorXf &p_vec, VectorXf &g_vec)
{
VectorXf d_vec(VectorXf::Map(d_cur.data(), d_cur.cols()*d_cur.rows()));
g_vec = L*p_vec - d_vec;
return (0.5*p_vec.transpose()*L*p_vec - d_vec.transpose()*p_vec)(0, 0);
}
int main(int argc, char *argv[])
{
perf_parse_args(argc, argv);
std::cout << "size: " << PERF_N << std::endl;
bolt::cl::control ctrl = bolt::cl::control::getDefault();
::cl::Device device = ctrl.getDevice();
std::cout << "device: " << device.getInfo<CL_DEVICE_NAME>() << std::endl;
// create device vector (filled with zeros)
bolt::cl::device_vector<int> d_vec(PERF_N, 0);
perf_timer t;
for(size_t trial = 0; trial < PERF_TRIALS; trial++){
t.start();
bolt::cl::fill(d_vec.begin(), d_vec.end(), int(trial));
t.stop();
}
std::cout << "time: " << t.min_time() / 1e6 << " ms" << std::endl;
return 0;
}
std::vector<Vector3 *> Scene::traceLine(Camera const *cam, Image const *img, int j) const
{
std::vector<Vector3 *> results(img->width());
Ray ray;
HitInfo hitInfo;
Vector3 shadeResult;
std::vector<float> refr_stack;
std::vector<Vector3> colors(m_samples);
for (int i = 0; i < img->width(); ++i)
{
for (int k = 0; k < m_samples; ++k)
{
ray = cam->eyeRay(i + (m_samples == 1 ? 0.5f : randone(g_rng) - 0.5f), j + (m_samples == 1 ? 0.5f : randone(g_rng) - 0.5f), img->width(), img->height());
if (m_focus_length >= 0.f) // if depth of field enabled
{
Vector3 focal_point = ray.o + m_focus_length * ray.d;
Vector3 view_dir = cam->viewDir();
Vector3 d_vec(randone(g_rng) - 0.5f, randone(g_rng) - 0.5f, randone(g_rng) - 0.5f);
d_vec.normalize();
d_vec = view_dir.cross(d_vec);
ray.o += d_vec * randone(g_rng) * m_lens;
ray.d = focal_point - ray.o;
ray.d.normalize();
}
refr_stack.clear();
refr_stack.push_back(1.f);
ray.refractionStack = &refr_stack;
ray.refractionIndex = 0;
if (!results[i])
results[i] = new Vector3();
if (trace(hitInfo, ray))
{
colors[k] = hitInfo.material->shade(ray, hitInfo, *this);
}
else
{
colors[k] = bgColor();
}
*results[i] += colors[k];
}
if (m_cutoffs > 0)
{
std::vector<int> biggest(m_cutoffs, -1);
for (int c = 0; c < m_cutoffs; ++c)
{
if (biggest[0] == -1 || colors[c].length2() > colors[biggest[0]].length2())
{
biggest[0] = c;
for (int k = 1; k < m_cutoffs; ++k)
{
if (biggest[k] == -1 || colors[biggest[k - 1]].length2() > colors[biggest[k]].length2())
std::swap(biggest[k - 1], biggest[k]);
else
break;
}
}
}
for (int k = 0; k < m_cutoffs; ++k)
*results[i] -= colors[biggest[k]];
}
if (results[i])
*results[i] /= m_samples - m_cutoffs;
}
return results;
}
