[ODE] ode & direct3d and double precision

[Pablo]

i have a problem and some question about ode.
the problem is that i use direct3d with x library and my applications crash
when i try to put the const dReal* position of a body in a d3dxvector 3
for set up a traslation matrix.

if pos is my pointer to const dReal then i write
D3DXVECTOR3 vectorpos = D3DXVECTOR3(pos[0],pos[1],pos[2]);
this crash the program.
I've also tryed to make a static cast call for turn the pos[] in a float
but don't work.

I have downloaded a just compiled ode.lib with double precision activated.
I want to make a game so Do i need this numeric precision???
can someone tell me where i can find a compiled ode for visual studio with
single precision?

I don't understand what integration ode use is it a sort of Verlet integration??

thanks.
excuse me for my bad english but i'm a joung italian guy.