StandardWellsDense: fix some valgrind complaints

note that I don't know if these changes are semantically correct (I
doubt it), but this patch fixes the valgrind complaints I saw for
SPE9_CP and on Norne.

Also, this makes the timing of flow_ebos for SPE9 determinisic between
runs: without this, I got some random time steps fail in a given run
and in the next run a completely different set of timesteps
failed. Since this was on the same same machine, without any
recompiles and no changes to the deck or any other input parameters, I
initially attributed the behavior to cosmic rays ;)

@totto82: could you have a thorough look on this?

opm/autodiff/BlackoilDetails.hpp

@@ -226,8 +226,7 @@ namespace detail {
             assert(nw * np == int(residual_well.size()));
 
             // Do the global reductions
-#if 0
-            HAVE_MPI
+#if 0 // HAVE_MPI
             if ( linsolver_.parallelInformation().type() == typeid(ParallelISTLInformation) )
             {
                 const ParallelISTLInformation& info =
@@ -276,7 +275,7 @@ namespace detail {
                 for ( int idx = 0; idx < np; ++idx )
                 {
                     B_avg[idx] = B.col(idx).sum()/nc;
-                    maxCoeff[idx] = tempV.col(idx).maxCoeff();
+                    //maxCoeff[idx] = tempV.col(idx).maxCoeff();
                     R_sum[idx] = R.col(idx).sum();
 
                     assert(np >= np);