Use mdapy efficiently

The key point is re-using the neighborlist information.

[1]:

import mdapy as mp
mp.init()

[Taichi] version 1.7.0, llvm 15.0.1, commit 2fd24490, win, python 3.8.0
[Taichi] Starting on arch=x64

[2]:

mp.__version__

[2]:

'0.10.2'

Build a FCC lattice with 4,000,000 atoms.

[3]:

FCC = mp.LatticeMaker(3.615, 'FCC', 100, 100, 100)
FCC.compute()

Generate a system class.

[4]:

system = mp.System(pos=FCC.pos, box=FCC.box, boundary=[1, 1, 1])

Check the system information.

[5]:

system

[5]:

Filename: None
Atom Number: 4000000
Simulation Box:
[[361.5   0.    0. ]
 [  0.  361.5   0. ]
 [  0.    0.  361.5]
 [  0.    0.    0. ]]
TimeStep: 0
Boundary: [1, 1, 1]
Particle Information:
shape: (4_000_000, 5)
┌─────────┬──────┬──────────┬──────────┬──────────┐
│ id      ┆ type ┆ x        ┆ y        ┆ z        │
│ ---     ┆ ---  ┆ ---      ┆ ---      ┆ ---      │
│ i32     ┆ i32  ┆ f64      ┆ f64      ┆ f64      │
╞═════════╪══════╪══════════╪══════════╪══════════╡
│ 1       ┆ 1    ┆ 0.0      ┆ 0.0      ┆ 0.0      │
│ 2       ┆ 1    ┆ 1.8075   ┆ 1.8075   ┆ 0.0      │
│ 3       ┆ 1    ┆ 1.8075   ┆ 0.0      ┆ 1.8075   │
│ 4       ┆ 1    ┆ 0.0      ┆ 1.8075   ┆ 1.8075   │
│ …       ┆ …    ┆ …        ┆ …        ┆ …        │
│ 3999997 ┆ 1    ┆ 357.885  ┆ 357.885  ┆ 357.885  │
│ 3999998 ┆ 1    ┆ 359.6925 ┆ 359.6925 ┆ 357.885  │
│ 3999999 ┆ 1    ┆ 359.6925 ┆ 357.885  ┆ 359.6925 │
│ 4000000 ┆ 1    ┆ 357.885  ┆ 359.6925 ┆ 359.6925 │
└─────────┴──────┴──────────┴──────────┴──────────┘

If we want to do a series of analysis, we can re-use the neighbor to save time.

Direct calculation.

[6]:

%%time
system.cal_centro_symmetry_parameter(12)
system.cal_ackland_jones_analysis()
system.cal_atomic_entropy()

CPU times: total: 2min 10s
Wall time: 8.39 s

[7]:

system.data.head()

[7]:

shape: (5, 8)

id	type	x	y	z	csp	aja	atomic_entropy
i32	i32	f64	f64	f64	f64	i32	f64
1	1	0.0	0.0	0.0	3.2952e-26	1	-8.645166
2	1	1.8075	1.8075	0.0	1.0984e-26	1	-8.645166
3	1	1.8075	0.0	1.8075	1.0984e-26	1	-8.645166
4	1	0.0	1.8075	1.8075	1.0984e-26	1	-8.645166
5	1	0.0	0.0	3.615	2.1968e-26	1	-8.645166

Re-use neighborlist information.

[8]:

%%time
system.build_neighbor(rc=5.0, max_neigh=50) # Obtain the neighbor first, the following calculation can use it.
system.cal_atomic_entropy()
system.cal_ackland_jones_analysis()
system.cal_centro_symmetry_parameter(12)

CPU times: total: 1min 36s
Wall time: 6.8 s

[9]:

system.data.head()

[9]:

shape: (5, 8)

id	type	x	y	z	csp	aja	atomic_entropy
i32	i32	f64	f64	f64	f64	i32	f64
1	1	0.0	0.0	0.0	3.2952e-26	1	-8.645166
2	1	1.8075	1.8075	0.0	1.0984e-26	1	-8.645166
3	1	1.8075	0.0	1.8075	1.0984e-26	1	-8.645166
4	1	0.0	1.8075	1.8075	1.0984e-26	1	-8.645166
5	1	0.0	0.0	3.615	2.1968e-26	1	-8.645166