diff --git a/amgprec/amg_s_matchboxp_mod.f90 b/amgprec/amg_s_matchboxp_mod.f90
index a7f41c24..04194836 100644
--- a/amgprec/amg_s_matchboxp_mod.f90
+++ b/amgprec/amg_s_matchboxp_mod.f90
@@ -272,9 +272,7 @@ contains
           write(0,*) 'Impossible: mate(k) > nc'
           cycle
         else
-
           if (ilaggr(k) == ilaggr_neginit) then
-
             wk   = w(k)
             widx = w(idx)
             wmax = max(abs(wk),abs(widx))
diff --git a/amgprec/impl/aggregator/processMatchedVertices.cpp b/amgprec/impl/aggregator/processMatchedVertices.cpp
index 2b2160e9..531c9d32 100644
--- a/amgprec/impl/aggregator/processMatchedVertices.cpp
+++ b/amgprec/impl/aggregator/processMatchedVertices.cpp
@@ -338,7 +338,7 @@ void processMatchedVerticesS(
 #ifdef COUNT_LOCAL_VERTEX
     MilanLongInt localVertices = 0;
 #endif
-    //#pragma omp parallel private(k, w, v, k1, adj1, adj2, adj11, adj12, ghostOwner, option) \
+#pragma omp parallel private(k, w, v, k1, adj1, adj2, adj11, adj12, ghostOwner, option) \
     firstprivate(privateU, StartIndex, EndIndex, privateQLocalVtx, privateQGhostVtx, \
        privateQMsgType, privateQOwner, UChunkBeingProcessed) \
        default(shared)  num_threads(NUM_THREAD)   \
diff --git a/amgprec/impl/aggregator/processMatchedVerticesAndSendMessages.cpp b/amgprec/impl/aggregator/processMatchedVerticesAndSendMessages.cpp
index 40d333a7..1631348d 100644
--- a/amgprec/impl/aggregator/processMatchedVerticesAndSendMessages.cpp
+++ b/amgprec/impl/aggregator/processMatchedVerticesAndSendMessages.cpp
@@ -26,7 +26,7 @@ void processMatchedVerticesAndSendMessagesD(
 					   vector<MilanLongInt> &QLocalVtx,
 					   vector<MilanLongInt> &QGhostVtx,
 					   vector<MilanLongInt> &QMsgType,
-					   vector<MilanInt> &QOwner,
+ 					   vector<MilanInt> &QOwner,
 					   vector<MilanLongInt> &privateQLocalVtx,
 					   vector<MilanLongInt> &privateQGhostVtx,
 					   vector<MilanLongInt> &privateQMsgType,
@@ -345,7 +345,7 @@ void processMatchedVerticesAndSendMessagesS(
 #ifdef COUNT_LOCAL_VERTEX
     MilanLongInt localVertices = 0;
 #endif
-    //#pragma omp parallel private(k, w, v, k1, adj1, adj2, adj11, adj12, ghostOwner, option) \
+#pragma omp parallel private(k, w, v, k1, adj1, adj2, adj11, adj12, ghostOwner, option) \
     firstprivate(Message, privateU, StartIndex, EndIndex, privateQLocalVtx, privateQGhostVtx,\
 		 privateQMsgType, privateQOwner, UChunkBeingProcessed) default(shared) \
         num_threads(NUM_THREAD)                                                        \
diff --git a/amgprec/impl/level/amg_d_base_onelev_memory_use.f90 b/amgprec/impl/level/amg_d_base_onelev_memory_use.f90
index 25534fd0..da56851c 100644
--- a/amgprec/impl/level/amg_d_base_onelev_memory_use.f90
+++ b/amgprec/impl/level/amg_d_base_onelev_memory_use.f90
@@ -98,8 +98,7 @@ subroutine amg_d_base_onelev_memory_use(lv,il,nl,ilmin,info,iout,verbosity,prefi
     prefix_ = ""
   end if
 
-  write(iout_,*) trim(prefix_)
-
+  if (me == 0)   write(iout_,*) trim(prefix_)
 
   if (global_) then
     allocate(sz(6))