Vector3iVector is 1000 times slower if provided with int64

### Checklist

- [X] I have searched for [similar issues](https://github.com/isl-org/Open3D/issues).
- [X] For Python issues, I have tested with the [latest development wheel](http://www.open3d.org/docs/latest/getting_started.html#development-version-pip).
- [X] I have checked the [release documentation](http://www.open3d.org/docs/release/) and the [latest documentation](http://www.open3d.org/docs/latest/) (for `master` branch).

### Proposed new feature or change

Vector3iVector is 1000 times slower if we provide an int64 array instead of an int32 which has a huge impact when casting or manipulating meshes if we're not careful. Automatically casting to int32, forbidding int64 or a warning message would help and prevent this from happening.

### References

_No response_

### Additional information

Example:
```
import trimesh
import open3d as o3d
import numpy as np

print(o3d.__version__)

mesh = trimesh.primitives.Sphere(radius=1, center=[0,0,0], subdivisions=8)
vertices = mesh.vertices.copy()
faces64 = mesh.faces.copy()
faces32 = faces64.astype(np.int32)
print(f"\nVertices {vertices.dtype}")
%time _ = o3d.utility.Vector3dVector(vertices)
print(f"\nFaces {faces64.dtype}")
%time _ = o3d.utility.Vector3iVector(faces64)
print(f"\nFaces {faces32.dtype}")
%time _ = o3d.utility.Vector3iVector(faces32)
```
Output:
```
0.16.1+1c84344a

Vertices float64
CPU times: user 22.6 ms, sys: 5.39 ms, total: 28 ms
Wall time: 1.74 ms

Faces int64
CPU times: user 1.42 s, sys: 313 ms, total: 1.73 s
Wall time: 1.37 s

Faces int32
CPU times: user 1.12 ms, sys: 268 µs, total: 1.39 ms
Wall time: 1.4 ms
```