diff --git a/openfpm_data b/openfpm_data
index 151921ecf7d72831724e2a2faed2c1a4ec1b8984..34d9f7e2d5816486416544727369312f77dad4d9 160000
--- a/openfpm_data
+++ b/openfpm_data
@@ -1 +1 @@
-Subproject commit 151921ecf7d72831724e2a2faed2c1a4ec1b8984
+Subproject commit 34d9f7e2d5816486416544727369312f77dad4d9
diff --git a/src/Vector/cuda/vector_dist_comm_util_funcs.cuh b/src/Vector/cuda/vector_dist_comm_util_funcs.cuh
index 1ae16b0d563a60a19de853058f069cc9af3723f5..4c1e87f4f02da17e2c293557b4f671b96632c6e1 100644
--- a/src/Vector/cuda/vector_dist_comm_util_funcs.cuh
+++ b/src/Vector/cuda/vector_dist_comm_util_funcs.cuh
@@ -8,6 +8,8 @@
 #ifndef VECTOR_DIST_COMM_UTIL_FUNCS_HPP_
 #define VECTOR_DIST_COMM_UTIL_FUNCS_HPP_
 
+#include "util/cuda/scan_ofp.cuh"
+
 #define SKIP_LABELLING 512
 #define KEEP_PROPERTIES 512
 
@@ -92,7 +94,7 @@ struct labelParticlesGhost_impl<dim,St,prop,Memory,layout_base,Decomposition,tru
 			// scan
 			//sc.scan_(proc_id_out,starts);
 			starts.resize(proc_id_out.size());
-			mgpu::scan((unsigned int *)proc_id_out.template getDeviceBuffer<0>(), proc_id_out.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
+			openfpm::scan((unsigned int *)proc_id_out.template getDeviceBuffer<0>(), proc_id_out.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
 			starts.template deviceToHost<0>(starts.size()-1,starts.size()-1);
 			size_t sz = starts.template get<0>(starts.size()-1);
 
@@ -265,7 +267,7 @@ struct local_ghost_from_dec_impl<dim,St,prop,Memory,layout_base,true>
 		box_f_dev.toKernel(),box_f_sv.toKernel(),v_pos.toKernel(),o_part_loc.toKernel(),g_m);
 
 		starts.resize(o_part_loc.size());
-		mgpu::scan((unsigned int *)o_part_loc.template getDeviceBuffer<0>(), o_part_loc.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
+		openfpm::scan((unsigned int *)o_part_loc.template getDeviceBuffer<0>(), o_part_loc.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
 
 		starts.template deviceToHost<0>(starts.size()-1,starts.size()-1);
 		size_t total = starts.template get<0>(starts.size()-1);
diff --git a/src/Vector/cuda/vector_dist_cuda_func_test.cu b/src/Vector/cuda/vector_dist_cuda_func_test.cu
index 8d49cd59b78ed35e22f6d78c31dff433addd95c3..799cc528f1ebc58444f1c60ffe51d77dfa3153e4 100644
--- a/src/Vector/cuda/vector_dist_cuda_func_test.cu
+++ b/src/Vector/cuda/vector_dist_cuda_func_test.cu
@@ -11,6 +11,7 @@
 #include "util/cuda/scan_cuda.cuh"
 #include "util/cuda/moderngpu/kernel_scan.hxx"
 #include "Vector/vector_dist.hpp"
+#include "util/cuda/scan_ofp.cuh"
 
 #define SUB_UNIT_FACTOR 1024
 
@@ -129,7 +130,7 @@ BOOST_AUTO_TEST_CASE( vector_ghost_process_local_particles )
 	starts.resize(o_part_loc.size());
 
 	auto & v_cl = create_vcluster();
-	mgpu::scan((unsigned int *)o_part_loc.template getDeviceBuffer<0>(), o_part_loc.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
+	openfpm::scan((unsigned int *)o_part_loc.template getDeviceBuffer<0>(), o_part_loc.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
 
 	starts.deviceToHost<0>(starts.size()-1,starts.size()-1);
 	size_t tot = starts.template get<0>(o_part_loc.size()-1);
diff --git a/src/Vector/vector_dist_comm.hpp b/src/Vector/vector_dist_comm.hpp
index 7c30dbd3c8fe31e1d9f9361244c6931c47bea48f..050fc8af2d2123d1072c2b19955efd81e1ddfc36 100644
--- a/src/Vector/vector_dist_comm.hpp
+++ b/src/Vector/vector_dist_comm.hpp
@@ -18,6 +18,7 @@
 
 #include "Vector/util/vector_dist_funcs.hpp"
 #include "cuda/vector_dist_comm_util_funcs.cuh"
+#include "util/cuda/scan_ofp.cuh"
 
 #define NO_POSITION 1
 #define WITH_POSITION 2
@@ -1139,7 +1140,7 @@ class vector_dist_comm
 			#else
 
 			starts.resize(v_cl.size());
-			mgpu::scan((unsigned int *)prc_sz.template getDeviceBuffer<0>(), prc_sz.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
+			openfpm::scan((unsigned int *)prc_sz.template getDeviceBuffer<0>(), prc_sz.size(), (unsigned int *)starts.template getDeviceBuffer<0>() , v_cl.getmgpuContext());
 
 			// move prc_sz to host
 			prc_sz.template deviceToHost<0>();