Input data is real, not complex

[mtazzari]

Input data is real, not complex

[mtazzari]

@fredRos what do you think? Feasible?

[fredRos]

yes, we should do that. I'm profiling the code right now and see a number of things we can improve. This one is high-level and makes perfect sense

[fredRos]

It is more work and requires more decisions:

• should fft2d accept a real image? It could not work in place anymore
• chi2 and sample don't provide access to the Fourier space image, so here it is easy to do
• could we fftshift the real image? We could save half of the memory transfer

[mtazzari]

A realistic use-case of galario employs only the sample() and/or the chi2() functions.
The other functions are for those who want to play with galario more in detail (like us) and I think it's fine to keep data as complex for all of them.

For sample() and chi2() functions it is easy to implement the change and I would start with them. In this way, internally nothing should change since we cast from dreal* to dcomplex* at the very beginning, before starting any operation.
For the CUDA version, I would do the casting after dreal* data has been copied to the GPU and then I would feed it to the dcomplex* data_d array that has been initialized with 0 imaginary part.

[fredRos]

What about shifting the real image? It seems to me like we could do that and only after the shift we'd add the imaginary part, perhaps on the device

[fredRos]

kernel unsigned int to float: https://stackoverflow.com/questions/9153861/typecasting-in-cuda-and-cublas

I'm following your suggestion now. I tried to do it in place but that would not allow multiple threads to operate concurrently. But then we have to use 50 % extra memory on the GPU to have a real and complex image until the complex image is properly constructed. This may be a problem for users with high-res images and small memory GPUs

Perhaps I can do the profiling to see if it's faster to do the construction on the CPU and then transfer. On all system I have seen so far it is safe to assume that there is more memory on the host available.

[fredRos]

fixed by #45