Mark Bates
Mark Bates
Yes, you can use the standard estimator.
Great. Yes that would be helpful - we could add it to the docs. Post back here if you need any more guidance. You might want to create a simple...
I think it should be something more like this. This really depends on how you are passing the data into Gpufit. Below, I assume that the data is interleaved -...
So, if your function is f(x,y,z) = (P0 * x + P1 * y, P0 * y + P2 * z) Then you just have to choose how you want...
If you have unequal numbers of X, Y, and Z values, then you would also have to specify the number of X, Y, and Z points to the model function,...
If your function lives in a three dimensional space, then the total data size must be N_x * N_y * N_z, where N_x is the number of X coordinates, etc....
The multi-dimensionality makes this "toy" model more complicated, but I think this is what you want. ``` __device__ void calculate_vector_function( float const * parameters, int const n_fits, int const n_points,...
Note that my modified example may not be the fastest solution - it is meant for illustrating the point. Also, the modulo operator is supposedly slow in CUDA.
In the short term, I agree it would make sense to include an example model in which the derivatives are computed numerically. Even if the model is trivial with analytic...
Regarding Jaxfit, I'd note that the package is designed to compute a single fit operation in each call, whereas Gpufit is designed for the parallel execution of large numbers of...