is there Matrix for Float but not Double?

Support
1

is there Matrix for Float but not Double?

2

Hello,

Yes, in C++ you can declare a double matrix (a 2D array). Just do, for example, a:

double myMatrix[10][10];

To have a 10x10 matrix of doubles.

take care,

PC

3

Hello @gpu

Your question concerns the VTK class vtkMatrix4x4 instead of the general concept of matrices in math/programming, correct?

If so, the answer is no, there isn’t. vtkMatrix4x4 internally uses double* and returns the data either as a double* or const double* as can be seen from the public API - double* vtkMatrix4x4::GetData() and const double* vtkMatrix4x4::GetData().

This requires VTK developers to make an additional copy to float* before uploading to OpenGL via glUniformMatrix* functions. vtkOpenGLUniforms::SetUniformMatrix(const char* name, vtkMatrix4x4* v) does a similar thing.

I believe the reason for using double precision in vtkMatrix4x4 is for better numeric tolerance when applying transformations to skewed meshes where the geometry has very large or very small bounds.

4

that’s what i want to avoid…seems have to write a custom class. Thanks!
moreover, is there API or common method which is the most effiect to do this copy?