Dear Colleague,

One bug in the Lattice Boltzmann (LBE) code and three bugs in the Dissipative 
Particle Dynamics (DPD) code of DL_MESO 2.4 have been discovered.

If producing output files in Legacy VTK format using the LBE code, incorrect 
values for densities of fluids numbered 1 and higher are being included. This 
can be fixed by replacing lbsy.nf in lines 248, 372, 452, 526, 630, 734, 803 and 
867 of lbpIOLegacyVTK.cpp with lbsy.nq, so that the lines now read:

  ipos[0] = float(fGetOneMassSite(&lbf[ilen*lbsitelength+iprop*lbsy.nq]));

In the DPD code, two issues were pointed out to us by Yaser Afshar, to whom we 
offer our thanks:

1. Initialization of systems from scratch may result in beads being positioned
   on top of or very close to each other: this may cause stability issues if 
   using hard sphere potentials, e.g. Lennard-Jones. This can be fixed by 
   inserting the following immediately after line 83 in start_module.f90:

     IF (ktype<2) nbds = 0

   Please note that this fix does not affect how molecules are inserted: if 
   hard-core beads are to be used in molecular systems, the use of CONFIG files 
   to give initial positions is highly recommended.

2. The potential energies for DPD interactions are incorrectly expressed for 
   systems with interaction lengths other than 1. This can be fixed by dividing 
   the current term by the interaction length, i.e. by changing lines 169, 430, 
   844, 1117 and 1391 in field_module.f90 and line 168 in manybody_module.f90 
   to:

     pot (i) = pot (i) + 0.5d0 * vk0 * (vk1 - rrr) * (vk1 - rrr) / vk1

   Note that this has no effect on calculated forces, so system trajectories are 
   not affected by this bug.

An additional bug was spotted in the DPD utilities for converting restart files 
when bonds are stored globally and therefore written to all export* files. This 
can be fixed by changing line 16 in exportconfig.f90 and line 20 in 
exportimage.f90 to:

  LOGICAL :: lexist, lgbnd

adding the following immediately after line 136 in exportconfig.f90 and line 125 
in exportimage.f90:

  ELSE IF (key (1:4) =="LGBN" .OR. key (1:4) =="lgbn") THEN
    word = getword (record, 2)
    IF (word==".true." .OR. word==".TRUE.") lgbnd = .true.
    IF (word==".false." .OR. word==".FALSE.") lgbnd = .false.

changing lines 188 in exportconfig.f90 and line 176 in exportimage.f90 to:

  IF (lgbnd) THEN
    numbonds = nbonds (j)
  ELSE
    numbonds = numbonds + nbonds (j)
  END IF

and inserting the following between lines 216 and 217 in exportconfig.f90 and 
lines 202 and 203 in exportimage.f90:

  IF (lgbnd) ibond = 0

All of the above bug fixes have been applied to the current DL_MESO release: a 
corrected version may be downloaded by registered users of version 2.4 without 
re-registering and decrypted using the same password.

Michael Seaton
2011-12-23