Hi
I am searching for a 3d-lib that help me simulate a real space environment. I need to image at least a satellite rotating earth and earth rotating sun.
But with a basic floating point data type of 32 bit (Irrlicht's f32) this seems impossible, cause it only covers a precision of 6 digits.
It seems impossible to position a starting rocket in a position 250 meters above the earth's surface, when the earth has an orbit of about 149,597,870,691 km arround the sun.
Is it possible to make Irrlicht _and_ one of the background libraries (OpenGl or DirectX) work with double precision (that is 64 bit floating point?
Regards
Ayman
need 64bit geometry and vector data
Well, the FPU precision of the GPU is up to the GPU, no way to change a GPU's floating point precision that I know of. DirectX, will force you into single point precision when creating a device, unless you pass in D3DCREATE_FPU_PRESERVE. You can change the FPU precision of your application, using the control87 functions.
As far as being able to handle 3D worlds that big, you'll need to scale it down and most likely break it up into sectors of some sort( quadtree, octtrees, portals, etc. )
As far as being able to handle 3D worlds that big, you'll need to scale it down and most likely break it up into sectors of some sort( quadtree, octtrees, portals, etc. )
