This MR adds
clFFT_GetSize(), which returns the estimated size of a plan similar to
cufftGetSize(). This is useful for comparing the measured memory usage with a memory model, e.g. GCT, ComputeFstat. Different to
cufftGetSize() we need also the
batchSize, as different to cuFFT for ECLFFT this is not set while the plan generation and is changeable in the execute step. But in our
lalsuite GPU functions, this is not changeable after plan generation.
@bema would assign this to you, but I havn't the rights to do this