I am not very clear to the meaning of "Max Mem / Proc" +in the "GPU Time Info (average)". +Is it the maximal of GPU memory used by one CPU core? +
+It is the maximum memory used at one time on the GPU for data storage by +a single MPI process. - Mike +
Hardware and software requirements:
To use this package, you currently need to have specific NVIDIA diff --git a/doc/Section_accelerate.txt b/doc/Section_accelerate.txt index bfa7efaa3d..400a4ca31b 100644 --- a/doc/Section_accelerate.txt +++ b/doc/Section_accelerate.txt @@ -348,7 +348,12 @@ NOTE: give command line switches in examples +I am not very clear to the meaning of "Max Mem / Proc" +in the "GPU Time Info (average)". +Is it the maximal of GPU memory used by one CPU core? +It is the maximum memory used at one time on the GPU for data storage by +a single MPI process. - Mike [Hardware and software requirements:] diff --git a/doc/pair_gayberne.html b/doc/pair_gayberne.html index 84241fcd9d..b41efdc59d 100644 --- a/doc/pair_gayberne.html +++ b/doc/pair_gayberne.html @@ -101,19 +101,19 @@ commands, they only need to be specified once for each atom type. non-zero, the three values are assigned to atom type I. If all the epsilon_i values are zero, they are ignored. If any of epsilon_j_a, epsilon_j_b, epsilon_j_c are non-zero, the three values are assigned -to atom type J. If all three epsilon_i values are zero, they are +to atom type J. If all three epsilon_j values are zero, they are ignored. Thus the typical way to define the epsilon_i and epsilon_j coefficients is to list their values in "pair_coeff I J" commands when I = J, but set them to 0.0 when I != J. If you do list them when I != J, you should insure they are consistent with their values in other -pair_coeff commands. +pair_coeff commands, since only the last setting will be in effect.
Note that if this potential is being used as a sub-style of pair_style hybrid, and there is no "pair_coeff I I" setting made for Gay-Berne for a particular type I (because I-I interactions are computed by another hybrid pair potential), then you still need to insure the epsilon a,b,c coefficients are assigned to -that type in a "pair_coeff I J" command. +that type. e.g. in a "pair_coeff I J" command.
IMPORTANT NOTE: If the epsilon a = b = c for an atom type, and if the shape of the particle itself is spherical, meaning its 3 shape diff --git a/doc/pair_gayberne.txt b/doc/pair_gayberne.txt index ed969b1af8..d7403df468 100755 --- a/doc/pair_gayberne.txt +++ b/doc/pair_gayberne.txt @@ -96,19 +96,19 @@ Specifically, if any of epsilon_i_a, epsilon_i_b, epsilon_i_c are non-zero, the three values are assigned to atom type I. If all the epsilon_i values are zero, they are ignored. If any of epsilon_j_a, epsilon_j_b, epsilon_j_c are non-zero, the three values are assigned -to atom type J. If all three epsilon_i values are zero, they are +to atom type J. If all three epsilon_j values are zero, they are ignored. Thus the typical way to define the epsilon_i and epsilon_j coefficients is to list their values in "pair_coeff I J" commands when I = J, but set them to 0.0 when I != J. If you do list them when I != J, you should insure they are consistent with their values in other -pair_coeff commands. +pair_coeff commands, since only the last setting will be in effect. Note that if this potential is being used as a sub-style of "pair_style hybrid"_pair_hybrid.html, and there is no "pair_coeff I I" setting made for Gay-Berne for a particular type I (because I-I interactions are computed by another hybrid pair potential), then you still need to insure the epsilon a,b,c coefficients are assigned to -that type in a "pair_coeff I J" command. +that type. e.g. in a "pair_coeff I J" command. IMPORTANT NOTE: If the epsilon a = b = c for an atom type, and if the shape of the particle itself is spherical, meaning its 3 shape diff --git a/doc/read_restart.html b/doc/read_restart.html index 9a993d5ef2..4bddba300e 100644 --- a/doc/read_restart.html +++ b/doc/read_restart.html @@ -109,6 +109,11 @@ restart files. The doc pages for individual pair styles note if this is the case. This is also true of bond_style hybrid (and angle_style, dihedral_style, improper_style hybrid).
+All settings made by the pair_modify command, +such as the shift and tail settings, are stored in the restart file +with the pair style. The one exception is the pair_modify +compute setting is not stored. +
Information about kspace_style settings are not stored in the restart file. Hence if you wish to use an Ewald or PPPM solver, these commands must be re-issued after the restart file is diff --git a/doc/read_restart.txt b/doc/read_restart.txt index b8fcc84d27..4626c12275 100644 --- a/doc/read_restart.txt +++ b/doc/read_restart.txt @@ -106,6 +106,11 @@ restart files. The doc pages for individual pair styles note if this is the case. This is also true of bond_style hybrid (and angle_style, dihedral_style, improper_style hybrid). +All settings made by the "pair_modify"_doc/pair_modify.html command, +such as the shift and tail settings, are stored in the restart file +with the pair style. The one exception is the "pair_modify +compute"_pair_modify.html setting is not stored. + Information about "kspace_style"_kspace_style.html settings are not stored in the restart file. Hence if you wish to use an Ewald or PPPM solver, these commands must be re-issued after the restart file is