The Atomic Simulation Environment Ask Hjorth Larsen and the ASE - - PowerPoint PPT Presentation

▶

Mar 10, 2024 357 likes •484 views

The Atomic Simulation Environment The Atomic Simulation Environment Ask Hjorth Larsen and the ASE development team Abinit developer meeting April 17, 2013 The Atomic Simulation Environment The atomic simulation environment Collection of

SLIDE 1

The Atomic Simulation Environment

Ask Hjorth Larsen and the ASE development team

Abinit developer meeting

April 17, 2013

SLIDE 2

The Atomic Simulation Environment

The atomic simulation environment

◮ Collection of tools for atomistic simulations ◮ ASE facilitates calculations through other codes ◮ Written in Python; used by writing Python scripts ◮ Free software. GNU LGPL v2.1+. ◮ https://wiki.fysik.dtu.dk/ase/ ◮ CAMPOS projects: https://wiki.fysik.dtu.dk/

Original citation

◮ S. R. Bahn and K. W. Jacobsen

An object-oriented scripting interface to a legacy electronic structure code

Comput. Sci. Eng., Vol. 4, 56-66, 2002

SLIDE 3

The Atomic Simulation Environment

Main functionality of ASE

◮ Defining structures (molecules, surfaces, ...) ◮ File I/O with many formats ◮ Algorithms for structure optimizations and dynamics ◮ Uniform calculator interfaces to external codes

Calculations as Python scripts

◮ Full programming environment ◮ Everything is scriptable ◮ Access to useful libraries (numpy, matplotlib, scipy, ...)

SLIDE 4

The Atomic Simulation Environment

Atoms Calculator External code generate input script retrieve and convert data "get potential energy" solve Schrödinger equation return the potential energy

Interface through file I/O

◮ ASE creates inputfile, runs

programme (see figure)

Calculator daemon

◮ Calculator runs in background ◮ Read/write using sockets

Direct linking

◮ Everything within one process

→ efficient and nice

◮ Also rather complicated

SLIDE 5

The Atomic Simulation Environment

High-level algorithms

◮ Structure optimization (BFGS, FIRE, MDMin,

minima hopping)

◮ Minimum-energy path (NEB) ◮ Molecular dynamics (Verlet, Nos´

e–Hoover, Langevin, Berendsen, ...)

◮ Vibrational analysis ◮ Electron transport

SLIDE 6

The Atomic Simulation Environment Atoms Calculator Dynamics Optimizer GPAW Dacapo Abinit ... Interface Implementations Vibrations BFGS MDMin MolecularDynamics Langevin Verlet NEB Constraints

SLIDE 7

The Atomic Simulation Environment

ASE calculators

Also: Gaussian, Mopac

SLIDE 8

The Atomic Simulation Environment

Demonstration: calculate energy and structure

from ase import Atoms from ase.optimize import BFGS from gpaw import GPAW from ase.units import Ry system = Atoms(’H2O’, positions =[[ -1 ,0 ,0] ,[1 ,0 ,0] ,[0 ,0 ,1]]) system.center(vacuum =3.0) calc = GPAW(mode=’lcao ’, basis=’dzp’, h=.2) system.set_calculator (calc) #E = system. get_potential_energy ()

pt = BFGS(system , trajectory=’opt2.traj ’)
pt.run(fmax =.05)

SLIDE 9

The Atomic Simulation Environment

Demonstration: Abinit through ASE

from ase import Atoms from ase.optimize import BFGS from ase.calculators.abinit import Abinit from ase.units import Ry system = Atoms(’H2O’, positions =[[ -1 ,0 ,0] ,[1 ,0 ,0] ,[0 ,0 ,1]]) system.center(vacuum =3.0) calc = Abinit(ecut =25 * Ry , toldff =1e-4) system.set_calculator (calc)

pt = BFGS(system , trajectory=’opt2.traj ’)
pt.run(fmax =.05)

SLIDE 10

The Atomic Simulation Environment

Bulk rutile

from ase.lattice.spacegroup import crystal a = 4.6 c = 2.95 rutile = crystal ([’Ti’, ’O’], basis =[(0, 0, 0), (0.3, 0.3, 0.0)] , spacegroup =136, cellpar =[a, a, c, 90, 90, 90])

SLIDE 11

The Atomic Simulation Environment

The ASE team—or rather parts of it

Andrew Peterson Anthony Goodrow Ask Hjorth Larsen Carsten Rostgaard Christian Glinsvad David Landis Elvar ¨ Orn J´

nsson

Felix Hanke George Tritsaris Heine Anton Hansen Ivano Castelli Jakob Blomquist Jakob Schiøtz Janne Blomqvist Janosch Michael Rauba Jens Jørgen Mortensen Jesper Friis Jesper Kleis Jingzhe Chen John Kitchin Jonas Bjork Jon B. Maronsson Jussi Enkovaara Karsten Wedel Jacobsen Kristen Kaasbjerg Lars Grabow Marcin Dulak Marco Vanin Markus Kaukonen Mattias Slabanja Michael Walter Mikkel Strange Poul Georg Moses Tao Jiang Thomas Olsen Troels Kofoed Jacobsen Lead developer: Jens Jørgen Mortensen Group leader: Karsten Wedel Jacobsen

SLIDE 12

The Atomic Simulation Environment

Conclusion

◮ Pluggable calculators ◮ Python scripting ◮ Structure generation ◮ Free / open source software

Links and info

◮ CAMPOS: https://wiki.fysik.dtu.dk/ ◮ ASE: https://wiki.fysik.dtu.dk/ase/ ◮ Mailing list: ase-users@listserv.fysik.dtu.dk ◮ IRC: #gpaw on irc.freenode.net ◮ Python: http://www.python.org/

The Atomic Simulation Environment

Ask Hjorth Larsen and the ASE development team

Abinit developer meeting

April 17, 2013

The atomic simulation environment

◮ Collection of tools for atomistic simulations ◮ ASE facilitates calculations through other codes ◮ Written in Python; used by writing Python scripts ◮ Free software. GNU LGPL v2.1+. ◮ https://wiki.fysik.dtu.dk/ase/ ◮ CAMPOS projects: https://wiki.fysik.dtu.dk/

Original citation

◮ S. R. Bahn and K. W. Jacobsen

An object-oriented scripting interface to a legacy electronic structure code

Main functionality of ASE

◮ Defining structures (molecules, surfaces, ...) ◮ File I/O with many formats ◮ Algorithms for structure optimizations and dynamics ◮ Uniform calculator interfaces to external codes

Calculations as Python scripts

◮ Full programming environment ◮ Everything is scriptable ◮ Access to useful libraries (numpy, matplotlib, scipy, ...)

Atoms Calculator External code generate input script retrieve and convert data "get potential energy" solve Schrödinger equation return the potential energy

Interface through file I/O

◮ ASE creates inputfile, runs

programme (see figure)

Calculator daemon

◮ Calculator runs in background ◮ Read/write using sockets

Direct linking

◮ Everything within one process

→ efficient and nice

◮ Also rather complicated

High-level algorithms

◮ Structure optimization (BFGS, FIRE, MDMin,

minima hopping)

◮ Minimum-energy path (NEB) ◮ Molecular dynamics (Verlet, Nos´

e–Hoover, Langevin, Berendsen, ...)

◮ Vibrational analysis ◮ Electron transport

ASE calculators

Also: Gaussian, Mopac

Demonstration: calculate energy and structure

Demonstration: Abinit through ASE

from ase import Atoms from ase.optimize import BFGS from ase.calculators.abinit import Abinit from ase.units import Ry system = Atoms(’H2O’, positions =[[ -1 ,0 ,0] ,[1 ,0 ,0] ,[0 ,0 ,1]]) system.center(vacuum =3.0) calc = Abinit(ecut =25 * Ry , toldff =1e-4) system.set_calculator (calc)

Bulk rutile

from ase.lattice.spacegroup import crystal a = 4.6 c = 2.95 rutile = crystal ([’Ti’, ’O’], basis =[(0, 0, 0), (0.3, 0.3, 0.0)] , spacegroup =136, cellpar =[a, a, c, 90, 90, 90])

The ASE team—or rather parts of it

Andrew Peterson Anthony Goodrow Ask Hjorth Larsen Carsten Rostgaard Christian Glinsvad David Landis Elvar ¨ Orn J´

Conclusion

◮ Pluggable calculators ◮ Python scripting ◮ Structure generation ◮ Free / open source software

Links and info

◮ CAMPOS: https://wiki.fysik.dtu.dk/ ◮ ASE: https://wiki.fysik.dtu.dk/ase/ ◮ Mailing list: ase-users@listserv.fysik.dtu.dk ◮ IRC: #gpaw on irc.freenode.net ◮ Python: http://www.python.org/

Thanks for listening