Move parallel primitives lib: moderngpu to CUB

src/Operators/Vector/vector_dist_operators_functions.hpp

@@ -286,7 +286,7 @@ struct point_scalar_process
 
 		auto & v_cl = create_vcluster<CudaMemory>();
 
-		openfpm::reduce((val_type *)ve.template getDeviceBuffer<0>(), ve.size(), (val_type *)(exp_tmp2[0].getDevicePointer()), mgpu::plus_t<val_type>(), v_cl.getmgpuContext());
+		openfpm::reduce((val_type *)ve.template getDeviceBuffer<0>(), ve.size(), (val_type *)(exp_tmp2[0].getDevicePointer()), gpu::plus_t<val_type>(), v_cl.getgpuContext());
 
 		exp_tmp2[0].deviceToHost();
 
@@ -325,8 +325,8 @@ struct point_scalar_process<val_type,is_sort,true>
 			openfpm::reduce(&((typename val_type::coord_type *)ve.template getDeviceBuffer<0>())[offset],
 						 ve.size(),
 						 (typename val_type::coord_type *)(exp_tmp2[0].getDevicePointer()),
-						 mgpu::plus_t<typename val_type::coord_type>(),
-						 v_cl.getmgpuContext());
+						 gpu::plus_t<typename val_type::coord_type>(),
+						 v_cl.getgpuContext());
 
 			exp_tmp2[0].deviceToHost();