Unverified Commit 83f5c331 authored by Steve Plimpton's avatar Steve Plimpton Committed by GitHub
Browse files

Merge pull request #886 from lammps/snap-quadratic 

updates to quadratic form of SNAP potential
parents d5ec7629 b5b70b42

doc/src/compute_sna_atom.txt

+13 −8
Original line number Diff line number Diff line
@@ -161,9 +161,9 @@ function. 


The keyword {bzeroflag} determines whether or not {B0}, the bispectrum

components of an atom with no neighbors, are subtracted from

the calculated bispectrum components. This optional keyword is only

available for compute {sna/atom}, as {snad/atom} and {snav/atom}

are unaffected by the removal of constant terms.

the calculated bispectrum components. This optional keyword 

normally only affects compute {sna/atom}. However, when

{quadraticflag} is on, it also affects {snad/atom} and {snav/atom}.



The keyword {quadraticflag} determines whether or not the

quadratic analogs to the bispectrum quantities are generated.
@@ -230,13 +230,18 @@ are 30, 90, and 180, respectively. With {quadratic} value=1, 
the numbers of columns are 930, 2790, and 5580, respectively.



If the {quadratic} keyword value is set to 1, then additional

columns are appended to each per-atom array, corresponding to

columns are generated, corresponding to

the products of all distinct pairs of  bispectrum components. If the

number of bispectrum components is {K}, then the number of distinct pairs

is  {K}({K}+1)/2. These are output in subblocks of  {K}({K}+1)/2 columns, using the same

ordering of sub-blocks as was used for the bispectrum

components. Within each sub-block, the ordering is upper-triangular,

(1,1),(1,2)...(1,{K}),(2,1)...({K}-1,{K}-1),({K}-1,{K}),({K},{K})

is  {K}({K}+1)/2.

For compute {sna/atom} these columns are appended to existing {K} columns.

The ordering of quadratic terms is upper-triangular,

(1,1),(1,2)...(1,{K}),(2,1)...({K}-1,{K}-1),({K}-1,{K}),({K},{K}).

For computes {snad/atom} and {snav/atom} each set of {K}({K}+1)/2

additional columns is inserted directly after each of sub-block

of linear terms i.e. linear and quadratic terms are contiguous.

So the nesting order from inside to outside is bispectrum component,

linear then quadratic, vector/tensor component, type.



These values can be accessed by any command that uses per-atom values

from a compute as input.  See "Section

examples/snap/log.20Apr18.snap.Mo_Chen.g++.1

0 → 100644
+127 −0
Original line number Diff line number Diff line 
LAMMPS (20 Apr 2018)

OMP_NUM_THREADS environment is not set. Defaulting to 1 thread. (../comm.cpp:90)

  using 1 OpenMP thread(s) per MPI task

# Demonstrate SNAP Ta potential



# Initialize simulation



variable nsteps index 100

variable nrep equal 4

variable a equal 3.160

units		metal



# generate the box and atom positions using a BCC lattice



variable nx equal ${nrep}

variable nx equal 4

variable ny equal ${nrep}

variable ny equal 4

variable nz equal ${nrep}

variable nz equal 4



boundary	p p p



lattice         bcc $a

lattice         bcc 3.16

Lattice spacing in x,y,z = 3.16 3.16 3.16

region		box block 0 ${nx} 0 ${ny} 0 ${nz}

region		box block 0 4 0 ${ny} 0 ${nz}

region		box block 0 4 0 4 0 ${nz}

region		box block 0 4 0 4 0 4

create_box	1 box

Created orthogonal box = (0 0 0) to (12.64 12.64 12.64)

  1 by 1 by 1 MPI processor grid

create_atoms	1 box

Created 128 atoms

  Time spent = 0.000223637 secs



mass 1 183.84



# choose potential



include Mo_Chen_PRM2017.snap



# DATE: 2017-09-18 CONTRIBUTOR: Chi Chen <chc273@eng.ucsd.edu> CITATION: C. Chen, Z. Deng, R. Tran, H. Tang, I.-H. Chu, S. P. Ong, "Accurate force field for molybdenum by machine learning large materials data" Physical Review Materials 1, 04 3603 (2017)

# Generated by Materials Virtual Lab

# Definition of SNAP potential.

pair_style snap

pair_coeff * * Mo_Chen_PRM2017.snapcoeff Mo Mo_Chen_PRM2017.snapparam Mo

Reading potential file Mo_Chen_PRM2017.snapcoeff with DATE: 2017-09-18

SNAP Element = Mo, Radius 0.5, Weight 1 

Reading potential file Mo_Chen_PRM2017.snapparam with DATE: 2017-09-18

SNAP keyword rcutfac 4.615858 

SNAP keyword twojmax 6 





# Setup output



thermo		10

thermo_modify norm yes



# Set up NVE run



timestep 0.5e-3

neighbor 1.0 bin

neigh_modify once no every 1 delay 0 check yes



# Run MD



velocity all create 300.0 4928459

fix 1 all nve

run             ${nsteps}

run             100

Neighbor list info ...

  update every 1 steps, delay 0 steps, check yes

  max neighbors/atom: 2000, page size: 100000

  master list distance cutoff = 5.61586

  ghost atom cutoff = 5.61586

  binsize = 2.80793, bins = 5 5 5

  1 neighbor lists, perpetual/occasional/extra = 1 0 0

  (1) pair snap, perpetual

      attributes: full, newton on

      pair build: full/bin/atomonly

      stencil: full/bin/3d

      bin: standard

Per MPI rank memory allocation (min/avg/max) = 3.507 | 3.507 | 3.507 Mbytes

Step Temp E_pair E_mol TotEng Press 

       0          300   -22.405975            0     -22.3675    2575.7657 

      10    294.77555   -22.405305            0     -22.3675    2756.6894 

      20    279.53011    -22.40335            0     -22.3675    3285.8272 

      30    255.52174    -22.40027            0     -22.3675    4122.8933 

      40     224.7299   -22.396321            0   -22.367499    5204.3499 

      50    189.67529   -22.391825            0   -22.367499    6449.1308 

      60    153.18862   -22.387145            0   -22.367499     7765.911 

      70    118.14998   -22.382652            0   -22.367499    9061.1616 

      80    87.224916   -22.378685            0   -22.367499     10247.68 

      90    62.623892    -22.37553            0   -22.367498    11250.067 

     100      45.9103   -22.373386            0   -22.367498    12011.726 

Loop time of 3.3917 on 1 procs for 100 steps with 128 atoms



Performance: 1.274 ns/day, 18.843 hours/ns, 29.484 timesteps/s

99.7% CPU use with 1 MPI tasks x 1 OpenMP threads



MPI task timing breakdown:

Section |  min time  |  avg time  |  max time  |%varavg| %total

---------------------------------------------------------------

Pair    | 3.3906     | 3.3906     | 3.3906     |   0.0 | 99.97

Neigh   | 0          | 0          | 0          |   0.0 |  0.00

Comm    | 0.00039721 | 0.00039721 | 0.00039721 |   0.0 |  0.01

Output  | 0.00023007 | 0.00023007 | 0.00023007 |   0.0 |  0.01

Modify  | 0.00021887 | 0.00021887 | 0.00021887 |   0.0 |  0.01

Other   |            | 0.0002868  |            |       |  0.01



Nlocal:    128 ave 128 max 128 min

Histogram: 1 0 0 0 0 0 0 0 0 0

Nghost:    727 ave 727 max 727 min

Histogram: 1 0 0 0 0 0 0 0 0 0

Neighs:    0 ave 0 max 0 min

Histogram: 1 0 0 0 0 0 0 0 0 0

FullNghs:  7424 ave 7424 max 7424 min

Histogram: 1 0 0 0 0 0 0 0 0 0



Total # of neighbors = 7424

Ave neighs/atom = 58

Neighbor list builds = 0

Dangerous builds = 0



Total wall time: 0:00:03

examples/snap/log.20Apr18.snap.Mo_Chen.g++.4

0 → 100644
+127 −0
Original line number Diff line number Diff line 
LAMMPS (20 Apr 2018)

OMP_NUM_THREADS environment is not set. Defaulting to 1 thread. (../comm.cpp:90)

  using 1 OpenMP thread(s) per MPI task

# Demonstrate SNAP Ta potential



# Initialize simulation



variable nsteps index 100

variable nrep equal 4

variable a equal 3.160

units		metal



# generate the box and atom positions using a BCC lattice



variable nx equal ${nrep}

variable nx equal 4

variable ny equal ${nrep}

variable ny equal 4

variable nz equal ${nrep}

variable nz equal 4



boundary	p p p



lattice         bcc $a

lattice         bcc 3.16

Lattice spacing in x,y,z = 3.16 3.16 3.16

region		box block 0 ${nx} 0 ${ny} 0 ${nz}

region		box block 0 4 0 ${ny} 0 ${nz}

region		box block 0 4 0 4 0 ${nz}

region		box block 0 4 0 4 0 4

create_box	1 box

Created orthogonal box = (0 0 0) to (12.64 12.64 12.64)

  1 by 2 by 2 MPI processor grid

create_atoms	1 box

Created 128 atoms

  Time spent = 0.000277281 secs



mass 1 183.84



# choose potential



include Mo_Chen_PRM2017.snap



# DATE: 2017-09-18 CONTRIBUTOR: Chi Chen <chc273@eng.ucsd.edu> CITATION: C. Chen, Z. Deng, R. Tran, H. Tang, I.-H. Chu, S. P. Ong, "Accurate force field for molybdenum by machine learning large materials data" Physical Review Materials 1, 04 3603 (2017)

# Generated by Materials Virtual Lab

# Definition of SNAP potential.

pair_style snap

pair_coeff * * Mo_Chen_PRM2017.snapcoeff Mo Mo_Chen_PRM2017.snapparam Mo

Reading potential file Mo_Chen_PRM2017.snapcoeff with DATE: 2017-09-18

SNAP Element = Mo, Radius 0.5, Weight 1 

Reading potential file Mo_Chen_PRM2017.snapparam with DATE: 2017-09-18

SNAP keyword rcutfac 4.615858 

SNAP keyword twojmax 6 





# Setup output



thermo		10

thermo_modify norm yes



# Set up NVE run



timestep 0.5e-3

neighbor 1.0 bin

neigh_modify once no every 1 delay 0 check yes



# Run MD



velocity all create 300.0 4928459

fix 1 all nve

run             ${nsteps}

run             100

Neighbor list info ...

  update every 1 steps, delay 0 steps, check yes

  max neighbors/atom: 2000, page size: 100000

  master list distance cutoff = 5.61586

  ghost atom cutoff = 5.61586

  binsize = 2.80793, bins = 5 5 5

  1 neighbor lists, perpetual/occasional/extra = 1 0 0

  (1) pair snap, perpetual

      attributes: full, newton on

      pair build: full/bin/atomonly

      stencil: full/bin/3d

      bin: standard

Per MPI rank memory allocation (min/avg/max) = 3.486 | 3.486 | 3.486 Mbytes

Step Temp E_pair E_mol TotEng Press 

       0          300   -22.405975            0     -22.3675    2575.7657 

      10    294.63153   -22.405286            0     -22.3675    2753.4662 

      20    278.98535    -22.40328            0     -22.3675     3272.416 

      30    254.38916   -22.400125            0     -22.3675    4091.8933 

      40    222.91191   -22.396088            0   -22.367499    5148.5505 

      50    187.16984   -22.391504            0   -22.367499    6362.2454 

      60    150.08253   -22.386747            0   -22.367499    7643.2732 

      70    114.60307   -22.382197            0   -22.367499    8900.2448 

      80    83.449257   -22.378201            0   -22.367499    10047.619 

      90    58.862643   -22.375048            0   -22.367498    11012.233 

     100     42.41931   -22.372939            0   -22.367498    11740.641 

Loop time of 1.91636 on 4 procs for 100 steps with 128 atoms



Performance: 2.254 ns/day, 10.646 hours/ns, 52.182 timesteps/s

97.9% CPU use with 4 MPI tasks x 1 OpenMP threads



MPI task timing breakdown:

Section |  min time  |  avg time  |  max time  |%varavg| %total

---------------------------------------------------------------

Pair    | 1.8147     | 1.8411     | 1.8875     |   2.1 | 96.07

Neigh   | 0          | 0          | 0          |   0.0 |  0.00

Comm    | 0.022276   | 0.069629   | 0.095057   |  10.7 |  3.63

Output  | 0.00032496 | 0.00065821 | 0.0016179  |   0.0 |  0.03

Modify  | 0.00019503 | 0.00020915 | 0.00023341 |   0.0 |  0.01

Other   |            | 0.00481    |            |       |  0.25



Nlocal:    32 ave 32 max 32 min

Histogram: 4 0 0 0 0 0 0 0 0 0

Nghost:    431 ave 431 max 431 min

Histogram: 4 0 0 0 0 0 0 0 0 0

Neighs:    0 ave 0 max 0 min

Histogram: 4 0 0 0 0 0 0 0 0 0

FullNghs:  1856 ave 1856 max 1856 min

Histogram: 4 0 0 0 0 0 0 0 0 0



Total # of neighbors = 7424

Ave neighs/atom = 58

Neighbor list builds = 0

Dangerous builds = 0



Total wall time: 0:00:01

examples/snap/log.5Oct16.snap.Ta06A.g++.1→examples/snap/log.20Apr18.snap.Ta06A.g++.1

+32 −19
Original line number Diff line number Diff line 
LAMMPS (5 Oct 2016)

LAMMPS (20 Apr 2018)

OMP_NUM_THREADS environment is not set. Defaulting to 1 thread. (../comm.cpp:90)

  using 1 OpenMP thread(s) per MPI task

# Demonstrate SNAP Ta potential



# Initialize simulation
@@ -31,12 +33,13 @@ Created orthogonal box = (0 0 0) to (13.264 13.264 13.264) 
  1 by 1 by 1 MPI processor grid

create_atoms	1 box

Created 128 atoms

  Time spent = 0.000328064 secs



mass 1 180.88



# choose potential



include Ta06A_pot.snap

include Ta06A.snap

# DATE: 2014-09-05 CONTRIBUTOR: Aidan Thompson athomps@sandia.gov CITATION: Thompson, Swiler, Trott, Foiles and Tucker, arxiv.org, 1409.3880 (2014)



# Definition of SNAP potential Ta_Cand06A
@@ -48,10 +51,9 @@ variable zblz equal 73 


# Specify hybrid with SNAP, ZBL



pair_style hybrid/overlay snap zbl ${zblcutinner} ${zblcutouter}

pair_style hybrid/overlay snap zbl 4 ${zblcutouter}

pair_style hybrid/overlay snap zbl 4 4.8



pair_style hybrid/overlay zbl ${zblcutinner} ${zblcutouter} snap

pair_style hybrid/overlay zbl 4 ${zblcutouter} snap

pair_style hybrid/overlay zbl 4 4.8 snap

pair_coeff 1 1 zbl ${zblz} ${zblz}

pair_coeff 1 1 zbl 73 ${zblz}

pair_coeff 1 1 zbl 73 73
@@ -61,10 +63,11 @@ SNAP Element = Ta, Radius 0.5, Weight 1 
Reading potential file Ta06A.snapparam with DATE: 2014-09-05

SNAP keyword rcutfac 4.67637 

SNAP keyword twojmax 6 

SNAP keyword gamma 1 

SNAP keyword rfac0 0.99363 

SNAP keyword rmin0 0 

SNAP keyword diagonalstyle 3 

SNAP keyword bzeroflag 0 

SNAP keyword quadraticflag 0 





# Setup output
@@ -85,13 +88,23 @@ fix 1 all nve 
run             ${nsteps}

run             100

Neighbor list info ...

  2 neighbor list requests

  update every 1 steps, delay 0 steps, check yes

  max neighbors/atom: 2000, page size: 100000

  master list distance cutoff = 5.8

  ghost atom cutoff = 5.8

  binsize = 2.9 -> bins = 5 5 5

Memory usage per processor = 2.92823 Mbytes

  binsize = 2.9, bins = 5 5 5

  2 neighbor lists, perpetual/occasional/extra = 2 0 0

  (1) pair zbl, perpetual, half/full from (2)

      attributes: half, newton on

      pair build: halffull/newton

      stencil: none

      bin: none

  (2) pair snap, perpetual

      attributes: full, newton on

      pair build: full/bin/atomonly

      stencil: full/bin/3d

      bin: standard

Per MPI rank memory allocation (min/avg/max) = 4.138 | 4.138 | 4.138 Mbytes

Step Temp E_pair E_mol TotEng Press 

       0          300    -11.85157            0   -11.813095    2717.1661 

      10    295.96579   -11.851053            0   -11.813095    2696.1559
@@ -104,20 +117,20 @@ Step Temp E_pair E_mol TotEng Press 
      80    124.04276   -11.829003            0   -11.813094     1537.703 

      90     97.37622   -11.825582            0   -11.813094    1734.9662 

     100    75.007873   -11.822714            0   -11.813094    1930.8005 

Loop time of 3.43062 on 1 procs for 100 steps with 128 atoms

Loop time of 2.53266 on 1 procs for 100 steps with 128 atoms



Performance: 1.259 ns/day, 19.059 hours/ns, 29.149 timesteps/s

99.9% CPU use with 1 MPI tasks x no OpenMP threads

Performance: 1.706 ns/day, 14.070 hours/ns, 39.484 timesteps/s

99.5% CPU use with 1 MPI tasks x 1 OpenMP threads



MPI task timing breakdown:

Section |  min time  |  avg time  |  max time  |%varavg| %total

---------------------------------------------------------------

Pair    | 3.4295     | 3.4295     | 3.4295     |   0.0 | 99.97

Pair    | 2.5313     | 2.5313     | 2.5313     |   0.0 | 99.95

Neigh   | 0          | 0          | 0          |   0.0 |  0.00

Comm    | 0.00043988 | 0.00043988 | 0.00043988 |   0.0 |  0.01

Output  | 0.00010014 | 0.00010014 | 0.00010014 |   0.0 |  0.00

Modify  | 0.00024533 | 0.00024533 | 0.00024533 |   0.0 |  0.01

Other   |            | 0.0002978  |            |       |  0.01

Comm    | 0.00051379 | 0.00051379 | 0.00051379 |   0.0 |  0.02

Output  | 0.00023317 | 0.00023317 | 0.00023317 |   0.0 |  0.01

Modify  | 0.00023675 | 0.00023675 | 0.00023675 |   0.0 |  0.01

Other   |            | 0.0003583  |            |       |  0.01



Nlocal:    128 ave 128 max 128 min

Histogram: 1 0 0 0 0 0 0 0 0 0
@@ -133,4 +146,4 @@ Ave neighs/atom = 58 
Neighbor list builds = 0

Dangerous builds = 0



Total wall time: 0:00:03
 
Total wall time: 0:00:02

examples/snap/log.5Oct16.snap.Ta06A.g++.4→examples/snap/log.20Apr18.snap.Ta06A.g++.4

+32 −19
Original line number Diff line number Diff line 
LAMMPS (5 Oct 2016)

LAMMPS (20 Apr 2018)

OMP_NUM_THREADS environment is not set. Defaulting to 1 thread. (../comm.cpp:90)

  using 1 OpenMP thread(s) per MPI task

# Demonstrate SNAP Ta potential



# Initialize simulation
@@ -31,12 +33,13 @@ Created orthogonal box = (0 0 0) to (13.264 13.264 13.264) 
  1 by 2 by 2 MPI processor grid

create_atoms	1 box

Created 128 atoms

  Time spent = 0.000288486 secs



mass 1 180.88



# choose potential



include Ta06A_pot.snap

include Ta06A.snap

# DATE: 2014-09-05 CONTRIBUTOR: Aidan Thompson athomps@sandia.gov CITATION: Thompson, Swiler, Trott, Foiles and Tucker, arxiv.org, 1409.3880 (2014)



# Definition of SNAP potential Ta_Cand06A
@@ -48,10 +51,9 @@ variable zblz equal 73 


# Specify hybrid with SNAP, ZBL



pair_style hybrid/overlay snap zbl ${zblcutinner} ${zblcutouter}

pair_style hybrid/overlay snap zbl 4 ${zblcutouter}

pair_style hybrid/overlay snap zbl 4 4.8



pair_style hybrid/overlay zbl ${zblcutinner} ${zblcutouter} snap

pair_style hybrid/overlay zbl 4 ${zblcutouter} snap

pair_style hybrid/overlay zbl 4 4.8 snap

pair_coeff 1 1 zbl ${zblz} ${zblz}

pair_coeff 1 1 zbl 73 ${zblz}

pair_coeff 1 1 zbl 73 73
@@ -61,10 +63,11 @@ SNAP Element = Ta, Radius 0.5, Weight 1 
Reading potential file Ta06A.snapparam with DATE: 2014-09-05

SNAP keyword rcutfac 4.67637 

SNAP keyword twojmax 6 

SNAP keyword gamma 1 

SNAP keyword rfac0 0.99363 

SNAP keyword rmin0 0 

SNAP keyword diagonalstyle 3 

SNAP keyword bzeroflag 0 

SNAP keyword quadraticflag 0 





# Setup output
@@ -85,13 +88,23 @@ fix 1 all nve 
run             ${nsteps}

run             100

Neighbor list info ...

  2 neighbor list requests

  update every 1 steps, delay 0 steps, check yes

  max neighbors/atom: 2000, page size: 100000

  master list distance cutoff = 5.8

  ghost atom cutoff = 5.8

  binsize = 2.9 -> bins = 5 5 5

Memory usage per processor = 2.91109 Mbytes

  binsize = 2.9, bins = 5 5 5

  2 neighbor lists, perpetual/occasional/extra = 2 0 0

  (1) pair zbl, perpetual, half/full from (2)

      attributes: half, newton on

      pair build: halffull/newton

      stencil: none

      bin: none

  (2) pair snap, perpetual

      attributes: full, newton on

      pair build: full/bin/atomonly

      stencil: full/bin/3d

      bin: standard

Per MPI rank memory allocation (min/avg/max) = 4.118 | 4.118 | 4.118 Mbytes

Step Temp E_pair E_mol TotEng Press 

       0          300    -11.85157            0   -11.813095    2717.1661 

      10     295.8664    -11.85104            0   -11.813095     2702.935
@@ -104,20 +117,20 @@ Step Temp E_pair E_mol TotEng Press 
      80    121.80051   -11.828715            0   -11.813094    1627.6911 

      90    95.262635   -11.825311            0   -11.813094    1812.9327 

     100    73.194645   -11.822481            0   -11.813094    1995.2199 

Loop time of 0.89193 on 4 procs for 100 steps with 128 atoms

Loop time of 1.3621 on 4 procs for 100 steps with 128 atoms



Performance: 4.843 ns/day, 4.955 hours/ns, 112.116 timesteps/s

99.9% CPU use with 4 MPI tasks x no OpenMP threads

Performance: 3.172 ns/day, 7.567 hours/ns, 73.416 timesteps/s

98.7% CPU use with 4 MPI tasks x 1 OpenMP threads



MPI task timing breakdown:

Section |  min time  |  avg time  |  max time  |%varavg| %total

---------------------------------------------------------------

Pair    | 0.84444    | 0.86772    | 0.88108    |   1.6 | 97.29

Pair    | 1.2867     | 1.309      | 1.35       |   2.1 | 96.10

Neigh   | 0          | 0          | 0          |   0.0 |  0.00

Comm    | 0.009577   | 0.023049   | 0.046417   |   9.8 |  2.58

Output  | 0.00024009 | 0.00026137 | 0.00027895 |   0.1 |  0.03

Modify  | 8.2493e-05 | 9.352e-05  | 0.00010061 |   0.1 |  0.01

Other   |            | 0.0008071  |            |       |  0.09

Comm    | 0.0096083  | 0.050652   | 0.072999   |  10.9 |  3.72

Output  | 0.00031447 | 0.00060236 | 0.0014303  |   0.0 |  0.04

Modify  | 0.00014234 | 0.00016212 | 0.00018811 |   0.0 |  0.01

Other   |            | 0.001728   |            |       |  0.13



Nlocal:    32 ave 32 max 32 min

Histogram: 4 0 0 0 0 0 0 0 0 0
@@ -133,4 +146,4 @@ Ave neighs/atom = 58 
Neighbor list builds = 0

Dangerous builds = 0



Total wall time: 0:00:00
 
Total wall time: 0:00:01

CRA Git | Maintained and supported by SUSTech CRA and CCSE