We also support CUDA.jl.
Before using a
CuArray simply invoke.
Our routines need as input array either only
CuArrays. To get the deconvolution running, both the PSF and the measured array needs to be a
CuArray. See also our 3D example here.
However, our approach to express the regularizers with Tullio.jl is currently not performant with GPUs. Therefore, to use
CuArrays with regularizers, you need to choose
TV_cuda. Other regularizers are not yet supported since we hope that Tullio.jl will be one day mature enough to produce reasonable fast gradients for CUDA kernels as well.