The DirectXMath library is built around single-precision floats. This is by far the most common data type to use in 3D graphics and games. Some games have added double-precision floats, although typically those are done as floating-point with a double-precision offset. The original xboxmath library didn't support doubles because the Xbox 360 PowerPC custom CPU was terrible at them.
The real challenge is that to be useful, the majority of the vector functions would need to be implemented to be double to be useful.
If I were to do this, I'd add it to the Extensions sub-project rather than putting it into the core library.
The DirectXMath library is built around single-precision floats. This is by far the most common data type to use in 3D graphics and games. Some games have added double-precision floats, although typically those are done as floating-point with a double-precision offset. The original xboxmath library didn't support doubles because the Xbox 360 PowerPC custom CPU was terrible at them.
The real challenge is that to be useful, the majority of the vector functions would need to be implemented to be double to be useful.
If I were to do this, I'd add it to the Extensions sub-project rather than putting it into the core library.