Opencl datahandling

Closes #15 (closed)

OpenCL kernels are now integrated in the normal create_kernel workflow. Also there exists a created a opencljit.init_globally function that just creates some CL queue/contex if you do not want to give it as a parameter to every kernel.

SerialDatahandling is extended to work with alternative GPU array libraries to PyCuda. There is now some overlapping code with the _custom_transfer_functions but I suppose they are for certain quantities that have a separate transfer function as oppose to using a whole different backend.

@kuron can you have a look on it? I think the solution is not as elegant as I thought it would be. pycuda.gpuarray.GPUArrays are not wrapped. So if you use dh.gpuarrays['foo'] you get either a pycuda array or a opencl array. I thought this step would be to drastic for one PR. Using OpenCL should still be a lot easier now.

pystencils/datahandling/__init__.py

@@ -20,7 +20,8 @@ def create_data_handling(domain_size: Tuple[int, ...],
                          default_layout: str = 'SoA',
                          default_target: str = 'cpu',
                          parallel: bool = False,
-                         default_ghost_layers: int = 1) -> DataHandling:
+                         default_ghost_layers: int = 1,
+                         opencl_queue=None) -> DataHandling:
     """Creates a data handling instance.
 
     Args:
@@ -33,6 +34,7 @@ def create_data_handling(domain_size: Tuple[int, ...],
         default_ghost_layers: default number of ghost layers if not overwritten in 'add_array'
     """
     if parallel:
+        assert not opencl_queue, "OpenCL is only supported for SerialDataHandling"
         if wlb is None:
             raise ValueError("Cannot create parallel data handling because walberla module is not available")
 
@@ -56,8 +58,12 @@ def create_data_handling(domain_size: Tuple[int, ...],
         return ParallelDataHandling(blocks=block_storage, dim=dim, default_target=default_target,
                                     default_layout=default_layout, default_ghost_layers=default_ghost_layers)
     else:
-        return SerialDataHandling(domain_size, periodicity=periodicity, default_target=default_target,
-                                  default_layout=default_layout, default_ghost_layers=default_ghost_layers)
+        return SerialDataHandling(domain_size,
+                                  periodicity=periodicity,
+                                  default_target=default_target,
+                                  default_layout=default_layout,
+                                  default_ghost_layers=default_ghost_layers,
+                                  opencl_queue=opencl_queue)
 
 
 __all__ = ['create_data_handling']