correct the description of how the F dot r contributions are computed

doc/src/compute_pressure.rst

@@ -46,13 +46,14 @@ system volume (or area in 2d).  The second term is the virial, equal to
 -dU/dV, computed for all pairwise as well as 2-body, 3-body, 4-body,
 many-body, and long-range interactions, where :math:`r_i` and
 :math:`f_i` are the position and force vector of atom *i*, and the black
-dot indicates a dot product.  When periodic boundary conditions are
+dot indicates a dot product.  This is computed in parallel for each
-used, N' necessarily includes periodic image (ghost) atoms outside the
+sub-domain and then summed over all parallel processes. Thus N'
-central box, and the position and force vectors of ghost atoms are thus
+necessarily includes atoms from neighboring sub-domains (so-called ghost
-included in the summation.  When periodic boundary conditions are not
+atoms) and the position and force vectors of ghost atoms are thus
-used, N' = N = the number of atoms in the system.  :doc:`Fixes <fix>`
+included in the summation.  Only when running in serial and without
-that impose constraints (e.g. the :doc:`fix shake <fix_shake>` command)
+periodic boundary conditions is N' = N = the number of atoms in the
-also contribute to the virial term.
+system.  :doc:`Fixes <fix>` that impose constraints (e.g. the :doc:`fix
+shake <fix_shake>` command) may also contribute to the virial term.
 
 A symmetric pressure tensor, stored as a 6-element vector, is also
 calculated by this compute.  The 6 components of the vector are