Dear Colleague,

Two bugs have been discovered in the Dissipative Particle Dynamics (DPD) code of 
DL_MESO 2.5, and there is news on a citable journal article on DL_MESO.

In systems initialised with low particles densities, it has been discovered that 
the construction of linked-cell lists occasionally produces infinite loops, 
leading to the code hanging during force calculations. To fix this bug, please 
add the following code after line 212 in domain_module.f90 and line 133 in 
domain_module_ser.f90, both inside the parlnk subroutines, to check the lists 
for infinite loops and break them open:

  DO icell = 1, nlx2*nly2*nlz2

    i = lct (icell)
    j = lct (icell)
    IF (i/=0) THEN
      safe = .true.
      DO WHILE (i/=0 .AND. j/=0)
        i = link (i)
        j = link (j)
        IF (j==0) EXIT
        j = link (j)
        IF (i==j) THEN
          safe = .false.
          EXIT
        END IF
      END DO
      IF (.NOT. safe) THEN
        i = lct (icell)
        DO WHILE (.true.)
          IF (link (i)==link (j)) THEN
            link (j) = 0
            EXIT
          ELSE
            i = link (i)
            j = link (j)
          END IF
        END DO
      END IF
    END IF

  END DO

For systems involving Lees-Edwards shearing boundaries, the application of 
pairwise thermostats between particles across these boundaries has been observed 
to produce non-linear velocity profiles, particularly when using high 
dissipative force parameters or thermostats that apply high viscosities, e.g. 
Lowe-Andersen. (Many thanks are due to Mary McDonald from The University of New 
South Wales, Sydney for bringing our attention to this bug.)

The proposed solution is to switch off pairwise thermostatting across shearing 
boundaries but still calculate interaction forces [ref: Chatterjee, Mol. Sim. 
33, 1233-1236 (2007)]. This can be implemented by minor changes to all force 
calculation subroutines in field_module.f90 (forces_* and dragforces_dpdvv). In 
each subroutine:

1. Change the line starting with LOGICAL to add ltherm as a new variable, e.g.
   line 41 in forces_mdvv:

     LOGICAL :: frzchk, ltherm

2. Insert the following lines immediately after the definitions of jx, jy and 
   jz, e.g. after line 84 in forces_mdvv:

     ltherm = (.NOT. (srftype==3 .AND. &
             &(srflgc (1) .AND. jx==0) .OR. (srflgc (2) .AND. jx==nlx+1) .OR. &
             &(srflgc (3) .AND. jy==0) .OR. (srflgc (4) .AND. jy==nly+1) .OR. &
             &(srflgc (5) .AND. jz==0) .OR. (srflgc (6) .AND. jz==nlz+1)))

3. Insert the following line immediately before the calculation of 
   thermostatting forces or inclusion of particle pairs in the thermostatting 
   list:

     IF (ltherm) THEN

   and the following line immediately afterwards:

     END IF

   i.e. around lines 187-194 for forces_mdvv, lines 454-461 for forces_dpdvv, 
   lines 1180-1191 for forces_peters and lines 1473-1499 for forces_stoyanov. 
   Alternatively for dragforces_dpdvv, change line 622 to

     IF (lct (jbox)>0 .AND. ltherm) THEN

   and for forces_lowe, change line 900 to

     IF (ltherm .AND. mtrnd (idnode)

Minor corrections have also been made to the user manual: this can be downloaded 
from the main DL_MESO website. The ShearFlow case study has also been updated to 
reflect the code changes made above, including updates to the Example 
Simulations page of the DL_MESO website.

We strongly recommend that all registered users should either make the code 
changes listed above or re-download DL_MESO using the instructions they received 
on registration to obtain the corrected version of the software. Re-registration 
is not required and the same password is used for decryption.

We are also delighted to report that a citable journal article about DL_MESO, 
"DL_MESO: highly scalable mesoscale simulations", has finally been published 
online by Molecular Simulation: it can be found at 
www.tandfonline.com/doi/full/10.1080/08927022.2013.772297 with links to download 
the full text as a PDF (although access to this is dependent on institution 
subscription). This article gives details of the DPD and Lattice Boltzmann 
Equation (LBE) codes available in the package, along with parallel scalability 
data for each code. It has yet to be assigned to an issue of the printed version 
of the journal (and is currently found among its "Latest articles"), but it can 
currently be cited using its DOI, i.e.:

M.A. Seaton, R.L. Anderson, S. Metz & W. Smith, "DL_MESO: highly scalable mesoscale simulations", Mol. Sim. (2013). doi:10.1080/08927022.2013.772297

We recommend that future publications that result from using DL_MESO should cite 
this paper as well as include the acknowledgement stated in section 4.2 of the 
Licence Agreement (which includes the link to the DL_MESO website).

Michael Seaton
2013-04-17