for (i = 0; i < N; i++) {
center[i] =
Point3( 2*drand48()-1, 2*drand48()-1, 2*drand48()-1);
radius[i] =
0.1 / pow(N, 0.333);
}
The camera parameters are as follows:
lookfrom = (0, 0, 2)
lookat (0,0,0)
fov = 90 degrees.
The scenes were traced at 512 x 512 and were sampled at 1 sample/pixal
( Here are the images )
| Number of Spheres | Time |
| 1 | 1.01s |
| 10 | 2.4s |
| 100 | 6.7s |
| 1000 | 18.44s |
| 10000 | 47.21s |
| 100000 | 119.93s |
| 1000000 | (yeah, yeah, I'm still tracin this one...) |
Conclusions:
Well, mine didn't seem to fair as well as some of the other students,
but I'm going attribute it to the fact that I plugged this into my big
general ray tracer, with a big ugly lighting equation and so forth. I'm
doing this only to make myself feel better, and not because any of the
above mentioned BS is actually the case (Although, there was a tremendous
improvement over my big general ray tracer without the bounding volume
hierarchy).