Added explanation of __cuda_callable__

 * For every function, EXPECT_EQ needs to be done, even for zeros in matrices.

 * Figure out __cuda_callable_. When trying to call __cuda_callable__ functions

 *      a segmentation fault (core dumped) is thrown.

 *  ==>__cuda_callable__ works only for CPU at the moment. (for loops vs thread kernel assignment)

 *  ==>__cuda_callable__ works only for CPU at the moment. (for loops vs thread kernel assignment).

 *                       If we want to use __cuda_callable__ on the GPU, we need to call it as a kernel.

 */

#include <TNL/Containers/Vector.h>