Some modifications have been made in different places so that GPU code runs a bit faster (>=10%).
This branch also includes average computation in GPU.
Finally, a new data structure has been created so that GPU arrays can all be passed under a single variable, reducing the number of variables passed to GPU kernels.
Some modifications have been made in different places so that GPU code runs a bit faster (>=10%).
This branch also includes average computation in GPU.
Finally, a new data structure has been created so that GPU arrays can all be passed under a single variable, reducing the number of variables passed to GPU kernels.