Merge pull request #1860 from akohlmey/collected-small-changes

Collected small changes for next patch release

doc/src/pair_modify.rst

@@ -13,18 +13,14 @@ Syntax
 
 * one or more keyword/value pairs may be listed
 * keyword = *pair* or *shift* or *mix* or *table* or *table/disp* or *tabinner*
-  or *tabinner/disp* or *tail* or *compute* or *nofdotr*
+  or *tabinner/disp* or *tail* or *compute* or *nofdotr* or *special* or 
+  *compute/tally*
   
   .. parsed-literal::
   
-       *pair* values = sub-style N *special* which wt1 wt2 wt3
+       *pair* value = sub-style N
-                    or sub-style N *compute/tally* flag
          sub-style = sub-style of :doc:`pair hybrid <pair_hybrid>`
-         N = which instance of sub-style (only if sub-style is used multiple times)
+         N = which instance of sub-style (1 to M), only specify if sub-style is used multiple times
-           *special* which wt1 wt2 wt3 = override *special_bonds* settings (optional)
-             which = *lj/coul* or *lj* or *coul*
-             w1,w2,w3 = 1-2, 1-3, and 1-4 weights from 0.0 to 1.0 inclusive
-           *compute/tally* flag = *yes* or *no*
        *mix* value = *geometric* or *arithmetic* or *sixthpower*
        *shift* value = *yes* or *no*
        *table* value = N
@@ -37,8 +33,11 @@ Syntax
          cutoff = inner cutoff at which to begin table (distance units)
        *tail* value = *yes* or *no*
        *compute* value = *yes* or *no*
-       *nofdotr*
+       *nofdotr* value = none
-
+       *special* values = which wt1 wt2 wt3
+          which = *lj/coul* or *lj* or *coul*
+          w1,w2,w3 = 1-2, 1-3, 1-4 weights from 0.0 to 1.0 inclusive
+       *compute/tally* value = *yes* or *no*
 
 
 Examples
@@ -53,47 +52,41 @@ Examples
    pair_modify pair lj/cut compute no
    pair_modify pair tersoff compute/tally no
    pair_modify pair lj/cut/coul/long 1 special lj/coul 0.0 0.0 0.0
+   pair_modify pair lj/cut/coul/long special lj 0.0 0.0 0.5 special coul 0.0 0.0 0.8333333
 
 Description
 """""""""""
 
-Modify the parameters of the currently defined pair style.  Not all
+Modify the parameters of the currently defined pair style.  If the
-parameters are relevant to all pair styles.
+pair style is :doc:`hybrid or hybrid/overlay <pair_hybrid>`, then the
+specified parameters are by default modified for all the hybrid sub-styles.
 
-If used, the *pair* keyword must appear first in the list of keywords.
+.. note::
-It can only be used with the :doc:`hybrid and hybrid/overlay <pair_hybrid>` pair styles.  It means that all the
-following parameters will only be modified for the specified
-sub-style.  If the sub-style is defined multiple times, then an
-additional numeric argument *N* must also be specified, which is a
-number from 1 to M where M is the number of times the sub-style was
-listed in the :doc:`pair_style hybrid <pair_hybrid>` command.  The extra
-number indicates which instance of the sub-style the remaining
-keywords will be applied to.  Note that if the *pair* keyword is not
-used, and the pair style is *hybrid* or *hybrid/overlay*\ , then all the
-specified keywords will be applied to all sub-styles.
 
-The *special* and *compute/tally* keywords can **only** be used in
+  The behavior for hybrid pair styles can be changed by using the *pair*
-conjunction with the *pair* keyword and must directly follow it.
+  keyword, which allows selection of a specific sub-style to apply all
-*special* allows to override the
+  remaining keywords to.
-:doc:`special_bonds <special_bonds>` settings for the specified sub-style.
+  The *special* and *compute/tally* keywords can **only** be
-*compute/tally* allows to disable or enable registering
+  used in conjunction with the *pair* keyword.  See further details about
-:doc:`compute \*/tally <compute_tally>` computes for a given sub-style.
+  these 3 keywords below.
-More details are given below.
 
 The *mix* keyword affects pair coefficients for interactions between
 atoms of type I and J, when I != J and the coefficients are not
 explicitly set in the input script.  Note that coefficients for I = J
 must be set explicitly, either in the input script via the
-"pair\_coeff" command or in the "Pair Coeffs" section of the :doc:`data file <read_data>`.  For some pair styles it is not necessary to
+:doc:`pair_coeff <pair_coeff>` command or in the "Pair Coeffs" section of the
-specify coefficients when I != J, since a "mixing" rule will create
+:doc:`data file <read_data>`.  For some pair styles it is not
-them from the I,I and J,J settings.  The pair\_modify *mix* value
+necessary to specify coefficients when I != J, since a "mixing" rule
-determines what formulas are used to compute the mixed coefficients.
+will create them from the I,I and J,J settings.  The pair\_modify
-In each case, the cutoff distance is mixed the same way as sigma.
+*mix* value determines what formulas are used to compute the mixed
+coefficients.  In each case, the cutoff distance is mixed the same way
+as sigma.
 
-Note that not all pair styles support mixing.  Also, some mix options
+Note that not all pair styles support mixing and some mix options
-are not available for certain pair styles.  See the doc page for
+are not available for certain pair styles. Also, there are additional
-individual pair styles for those restrictions.  Note also that the
+restrictions when using :doc:`pair style hybrid or hybrid/overlay <pair_hybrid>`.
-:doc:`pair_coeff <pair_coeff>` command also can be to directly set
+See the doc page for individual pair styles for those restrictions.  Note also that the
+:doc:`pair_coeff <pair_coeff>` command also can be used to directly set
 coefficients for a specific I != J pairing, in which case no mixing is
 performed.
 
@@ -135,10 +128,11 @@ see the doc page for individual styles to see which potentials support
 these options.  If N is non-zero, a table of length 2\^N is
 pre-computed for forces and energies, which can shrink their
 computational cost by up to a factor of 2.  The table is indexed via a
-bit-mapping technique :ref:`(Wolff) <Wolff1>` and a linear interpolation is
+bit-mapping technique :ref:`(Wolff) <Wolff1>` and a linear
-performed between adjacent table values.  In our experiments with
+interpolation is performed between adjacent table values.  In our
-different table styles (lookup, linear, spline), this method typically
+experiments with different table styles (lookup, linear, spline), this
-gave the best performance in terms of speed and accuracy.
+method typically gave the best performance in terms of speed and
+accuracy.
 
 The choice of table length is a tradeoff in accuracy versus speed.  A
 larger N yields more accurate force computations, but requires more
@@ -162,27 +156,28 @@ pairwise interactions are computed via table lookup for simulations
 with "real" units, but some close pairs may be computed directly
 (non-table) for simulations with "lj" units.
 
-When the *tail* keyword is set to *yes*\ , certain pair styles will add
+When the *tail* keyword is set to *yes*\ , certain pair styles will
-a long-range VanderWaals tail "correction" to the energy and pressure.
+add a long-range VanderWaals tail "correction" to the energy and
-These corrections are bookkeeping terms which do not affect dynamics,
+pressure.  These corrections are bookkeeping terms which do not affect
-unless a constant-pressure simulation is being performed.  See the doc
+dynamics, unless a constant-pressure simulation is being performed.
-page for individual styles to see which support this option.  These
+See the doc page for individual styles to see which support this
-corrections are included in the calculation and printing of
+option.  These corrections are included in the calculation and
-thermodynamic quantities (see the :doc:`thermo_style <thermo_style>`
+printing of thermodynamic quantities (see the :doc:`thermo_style
-command).  Their effect will also be included in constant NPT or NPH
+<thermo_style>` command).  Their effect will also be included in
-simulations where the pressure influences the simulation box
+constant NPT or NPH simulations where the pressure influences the
-dimensions (e.g. the :doc:`fix npt <fix_nh>` and :doc:`fix nph <fix_nh>`
+simulation box dimensions (e.g. the :doc:`fix npt <fix_nh>` and
-commands).  The formulas used for the long-range corrections come from
+:doc:`fix nph <fix_nh>` commands).  The formulas used for the
-equation 5 of :ref:`(Sun) <Sun>`.
+long-range corrections come from equation 5 of :ref:`(Sun) <Sun>`.
 
 .. note::
 
    The tail correction terms are computed at the beginning of each
    run, using the current atom counts of each atom type.  If atoms are
-   deleted (or lost) or created during a simulation, e.g. via the :doc:`fix gcmc <fix_gcmc>` command, the correction factors are not
+   deleted (or lost) or created during a simulation, e.g. via the
-   re-computed.  If you expect the counts to change dramatically, you can
+   :doc:`fix gcmc <fix_gcmc>` command, the correction factors are not
-   break a run into a series of shorter runs so that the correction
+   re-computed.  If you expect the counts to change dramatically, you
-   factors are re-computed more frequently.
+   can break a run into a series of shorter runs so that the
+   correction factors are re-computed more frequently.
 
 Several additional assumptions are inherent in using tail corrections,
 including the following:
@@ -191,26 +186,27 @@ including the following:
   should not be used for systems that are non-liquid, 2d, have a slab
   geometry (only 2d periodic), or inhomogeneous.
 * G(r), the radial distribution function (rdf), is unity beyond the
-  cutoff, so a fairly large cutoff should be used (i.e. 2.5 sigma for an
+  cutoff, so a fairly large cutoff should be used (i.e. 2.5 sigma for
-  LJ fluid), and it is probably a good idea to verify this assumption by
+  an LJ fluid), and it is probably a good idea to verify this
-  checking the rdf.  The rdf is not exactly unity beyond the cutoff for
+  assumption by checking the rdf.  The rdf is not exactly unity beyond
-  each pair of interaction types, so the tail correction is necessarily
+  the cutoff for each pair of interaction types, so the tail
-  an approximation.
+  correction is necessarily an approximation.
 
-  The tail corrections are computed at the beginning of each simulation
+  The tail corrections are computed at the beginning of each
-  run.  If the number of atoms changes during the run, e.g. due to atoms
+  simulation run.  If the number of atoms changes during the run,
-  leaving the simulation domain, or use of the :doc:`fix gcmc <fix_gcmc>`
+  e.g. due to atoms leaving the simulation domain, or use of the
-  command, then the corrections are not updated to reflect the changed
+  :doc:`fix gcmc <fix_gcmc>` command, then the corrections are not
-  atom count.  If this is a large effect in your simulation, you should
+  updated to reflect the changed atom count.  If this is a large
-  break the long run into several short runs, so that the correction
+  effect in your simulation, you should break the long run into
-  factors are re-computed multiple times.
+  several short runs, so that the correction factors are re-computed
+  multiple times.
 
-* Thermophysical properties obtained from calculations with this option
+* Thermophysical properties obtained from calculations with this
-  enabled will not be thermodynamically consistent with the truncated
+  option enabled will not be thermodynamically consistent with the
-  force-field that was used.  In other words, atoms do not feel any LJ
+  truncated force-field that was used.  In other words, atoms do not
-  pair interactions beyond the cutoff, but the energy and pressure
+  feel any LJ pair interactions beyond the cutoff, but the energy and
-  reported by the simulation include an estimated contribution from
+  pressure reported by the simulation include an estimated
-  those interactions.
+  contribution from those interactions.
 
 
 The *compute* keyword allows pairwise computations to be turned off,
@@ -225,53 +221,73 @@ a simulation with :doc:`pair_style hybrid <pair_hybrid>` with only a
 subset of the hybrid sub-styles enabled.  Second, this option allows
 you to perform a simulation with only long-range interactions but no
 short-range pairwise interactions.  Doing this by simply not defining
-a pair style will not work, because the
+a pair style will not work, because the :doc:`kspace_style
-:doc:`kspace_style <kspace_style>` command requires a Kspace-compatible
+<kspace_style>` command requires a Kspace-compatible pair style be
-pair style be defined.
+defined.
 
 The *nofdotr* keyword allows to disable an optimization that computes
-the global stress tensor from the total forces and atom positions rather
+the global stress tensor from the total forces and atom positions
-than from summing forces between individual pairs of atoms.
+rather than from summing forces between individual pairs of atoms.
 
 
 ----------
 
+The *pair* keyword can only be used with the :doc:`hybrid and
+hybrid/overlay <pair_hybrid>` pair styles.  If used, it must appear
+first in the list of keywords.
 
-The *special* keyword allows to override the 1-2, 1-3, and 1-4
+Its meaning is that all the following parameters will only be modified
-exclusion settings for individual sub-styles of a
+for the specified sub-style.  If the sub-style is defined multiple
-:doc:`hybrid pair style <pair_hybrid>`. It requires 4 arguments similar
+times, then an additional numeric argument *N* must also be specified,
-to the :doc:`special_bonds <special_bonds>` command, *which* and
+which is a number from 1 to M where M is the number of times the
-wt1,wt2,wt3.  The *which* argument can be *lj* to change the
+sub-style was listed in the :doc:`pair_style hybrid <pair_hybrid>`
-Lennard-Jones settings, *coul* to change the Coulombic settings,
+command.  The extra number indicates which instance of the sub-style
-or *lj/coul* to change both to the same set of 3 values.  The wt1,wt2,wt3
+the remaining keywords will be applied to.
-values are numeric weights from 0.0 to 1.0 inclusive, for the 1-2,
+
-1-3, and 1-4 bond topology neighbors, respectively. The *special*
+The *special* and *compute/tally* keywords can **only** be used in
-keyword can only be used in conjunction with the *pair* keyword
+conjunction with the *pair* keyword and they must directly follow it.
-and has to directly follow it.  This option is not compatible with
+I.e. any other keyword, must appear after *pair*, *special*, and
-pair styles from the GPU or the USER-INTEL package and attempting
+*compute/tally*.
-it will cause an error.
+
+The *special* keyword overrides the global :doc:`special_bonds <special_bonds>`
+1-2, 1-3, 1-4 exclusion settings (weights) for the sub-style selected
+by the *pair* keyword.
+
+Similar to the :doc:`special_bonds <special_bonds>` command, it takes
+4 arguments.  The *which* argument can be *lj* to change only the
+non-Coulomb weights (e.g. Lennard-Jones or Buckingham), *coul* to change
+only the Coulombic settings, or *lj/coul* to change both to the same
+values.  The *wt1,wt2,wt3* values are numeric weights from 0.0 to 1.0
+inclusive, for the 1-2, 1-3, and 1-4 bond topology neighbors, respectively.
+The *special* keyword can be used multiple times, e.g. to set the *lj*
+and *coul* settings to different values.
 
 .. note::
 
-   The global settings specified by the
+   The *special* keyword is not compatible with pair styles from the
-   :doc:`special_bonds <special_bonds>` command affect the construction of
+   GPU or the USER-INTEL package and attempting to use it will cause
-   neighbor lists.  Weights of 0.0 (for 1-2, 1-3, or 1-4 neighbors)
+   an error.
-   exclude those pairs from the neighbor list entirely.  Weights of 1.0
-   store the neighbor with no weighting applied. Thus only global values
-   different from exactly 0.0 or 1.0 can be overridden and an error is
-   generated if the requested setting is not compatible with the global
-   setting. Substituting 1.0e-10 for 0.0 and 0.9999999999 for 1.0 is
-   usually a sufficient workaround in this case without causing a
-   significant error.
-
-The *compute/tally* keyword takes exactly 1 argument (\ *no* or *yes*\ ),
-and allows to selectively disable or enable processing of the various
-:doc:`compute \*/tally <compute_tally>` styles for a given
-:doc:`pair hybrid or hybrid/overlay <pair_hybrid>` sub-style.
 
 .. note::
 
-   Any "pair\_modify pair compute/tally" command must be issued
+   Weights of exactly 0.0 or 1.0 in the :doc:`special_bonds <special_bonds>`
+   command have implications on the neighbor list construction, which
+   means that they cannot be overridden by using the *special* keyword.
+   One workaround for this restriction is to use the :doc:`special_bonds <special_bonds>`
+   command with weights like 1.0e-10 or 0.999999999 instead of 0.0 or 1.0,
+   respectively, which enables to reset each them to any value between 0.0
+   and 1.0 inclusively.  Otherwise you can set **all** global weights to
+   an arbitrary number between 0.0 or 1.0, like 0.5, and then you have
+   to override **all** *special* settings for **all** sub-styles which use
+   the 1-2, 1-3, and 1-4 exclusion weights in their force/energy computation.
+
+The *compute/tally* keyword disables or enables registering :doc:`compute
+\*/tally <compute_tally>` computes for the sub-style specified by
+the *pair* keyword.  Use *no* to disable, or *yes* to enable.
+
+.. note::
+
+   The "pair_modify pair compute/tally" command must be issued
    **before** the corresponding compute style is defined.
 
 
@@ -290,7 +306,7 @@ Related commands
 """"""""""""""""
 
 :doc:`pair_style <pair_style>`, :doc:`pair_style hybrid <pair_hybrid>`,
-pair\_coeff"_pair\_coeff.html, :doc:`thermo_style <thermo_style>`,
+:doc:`pair_coeff <pair_coeff>`, :doc:`thermo_style <thermo_style>`,
 :doc:`compute \*/tally <compute_tally>`
 
 Default

doc/src/set.rst

@@ -359,9 +359,9 @@ particles must be ellipsoids as defined by the :doc:`atom_style ellipsoid <atom_
 of the particles is set to the 3 specified components.
 
 Keyword *omega* sets the angular velocity of selected atoms.  The
-particles must be spheres as defined by the "atom\_style sphere"_
+particles must be spheres as defined by the
-atom\_style.html command.  The angular velocity vector of the particles
+:doc:`atom_style sphere <atom_style>` command.  The angular velocity
-is set to the 3 specified components.
+vector of the particles is set to the 3 specified components.
 
 Keyword *mass* sets the mass of all selected particles.  The particles
 must have a per-atom mass attribute, as defined by the

src/kspace.cpp

@@ -286,7 +286,7 @@ void KSpace::qsum_qsq(int warning_flag)
   double qsum_local(0.0), qsqsum_local(0.0);
 
 #if defined(_OPENMP)
-#pragma omp parallel for default(none) reduction(+:qsum_local,qsqsum_local)
+#pragma omp parallel for default(shared) reduction(+:qsum_local,qsqsum_local)
 #endif
   for (int i = 0; i < nlocal; i++) {
     qsum_local += q[i];

src/neighbor.cpp

@@ -402,13 +402,6 @@ void Neighbor::init()
     }
   }
 
-  // maxwt = max multiplicative factor on atom indices stored in neigh list
-
-  maxwt = 0;
-  if (special_flag[1] == 2) maxwt = 2;
-  if (special_flag[2] == 2) maxwt = 3;
-  if (special_flag[3] == 2) maxwt = 4;
-
   // ------------------------------------------------------------------
   // xhold array
 

src/neighbor.h

@@ -166,7 +166,6 @@ class Neighbor : protected Pointers {
 
   int maxatom;                     // max size of atom-based NeighList arrays
   int maxrequest;                  // max size of NeighRequest list
-  int maxwt;                       // max weighting factor applied + 1
 
   // info for other Neigh classes: NBin,NStencil,NPair,NTopo
 

src/pair_hybrid.cpp

@@ -859,14 +859,17 @@ void PairHybrid::modify_params(int narg, char **arg)
       iarg = 3;
     }
 
-    // if 2nd keyword (after pair) is special:
+    // keywords "special" and "compute/tally" have to be listed directly
-    // invoke modify_special() for the sub-style
+    // after "pair" but can be given multiple times
+
+again:
 
     if (iarg < narg && strcmp(arg[iarg],"special") == 0) {
       if (narg < iarg+5)
         error->all(FLERR,"Illegal pair_modify special command");
       modify_special(m,narg-iarg,&arg[iarg+1]);
       iarg += 5;
+      goto again;
     }
 
     // if 2nd keyword (after pair) is compute/tally:
@@ -881,11 +884,13 @@ void PairHybrid::modify_params(int narg, char **arg)
         compute_tally[m] = 0;
       } else error->all(FLERR,"Illegal pair_modify compute/tally command");
       iarg += 2;
+      goto again;
     }
 
-    // apply the remaining keywords to the base pair style itself and the
+    // apply the remaining keywords to the base pair style itself and
-    // sub-style except for "pair" and "special".
+    // the sub-style except for "pair" and "special" or "compute/tally"
-    // the former is important for some keywords like "tail" or "compute"
+    // and their arguments. the former is important for some keywords
+    // like "tail" or "compute"
 
     if (narg-iarg > 0) {
       Pair::modify_params(narg-iarg,&arg[iarg]);

src/random_mars.cpp

@@ -183,7 +183,7 @@ void RanMars::select_subset(bigint ntarget, int nmine, int *mark, int *next)
   int mode,index,oldindex,newvalue,nflip,which,niter;
   int active[2],first[2],last[2];
   int newactive[2],newfirst[2],newlast[2];
-  bigint nmark,nactive,nactiveall,nflipall,bnflip;
+  bigint nmark,nflipall;
   bigint activeall[2],bsum[3],bsumall[3];
   double thresh;
 
@@ -210,15 +210,12 @@ void RanMars::select_subset(bigint ntarget, int nmine, int *mark, int *next)
 
     // choose to ADD or SUBTRACT from current nmark
     // thresh = desired flips / size of active set
-    // nactive = size of current active set, only for debug output below
 
     if (ntarget-nmark > 0) {
       mode = ADD;
-      // nactive = active[mode];
       thresh = 1.0 * (ntarget-nmark) / activeall[mode];
     } else {
       mode = SUBTRACT;
-      // nactive = active[mode];
       thresh = 1.0 * (nmark-ntarget) / activeall[mode];
     }
 
@@ -278,12 +275,10 @@ void RanMars::select_subset(bigint ntarget, int nmine, int *mark, int *next)
     bsum[0] = nflip;
     bsum[1] = active[0];
     bsum[2] = active[1];
-    bsum[3] = nactive;
+    MPI_Allreduce(&bsum,&bsumall,3,MPI_LMP_BIGINT,MPI_SUM,world);
-    MPI_Allreduce(&bsum,&bsumall,4,MPI_LMP_BIGINT,MPI_SUM,world);
     nflipall = bsumall[0];
     activeall[0] = bsumall[1];
     activeall[1] = bsumall[2];
-    nactiveall = bsumall[3];
 
     if (mode == ADD) nmark += nflipall;
     else if (mode == SUBTRACT) nmark -= nflipall;