GPUMD-v2.8.1

• In GPUMD-v2.8, the regression accuracy of the NEP potential for single-component system has been decreased compared to the previous versions, due a change related to using mini-batch for training. Here in GPUMD-v2.8.1, we bring back the original accuracy.

GPUMD-v2.8

• Enabled using a batchsize in NEP training (Therefore, one more line is added to nep.in).
• Now NEP needs both train.in and test.in and will output energy_train.out, force_train.out, virial_train.out, energy_test.out, force_test.out, and virial_test.out instead of energy.out, force.out, and virial.out only.
• Now for the NEP potential, the atom types in train.in, test.in (for the nep executable), and xyz.in (for the gpumd and phonon executable) are not represented as integers but as atom symbols (such as H, He, and Li).
• The NEP training efficiency has been significantly improved when there are a lot of small-cell structures.
• Fixed a bug related to thermo.out (the triclinic box matrix was different from the manual by a transpose).
• Fixed a bug for the dump_position, dump_velocity, and dump_force keywords (These keywords may output incorrect data when the group option is used and the group indices in the relevant grouping method are not continuous).

GPUMD-v2.7

Major changes compared to GPUMD-v2.6:

• Improved the accuracy of the NEP potential for multi-component systems. For details, see Ref. [1].
• The formats of the input files related to the NEP potential have been slightly changed. To use the original NEP potential as described in Ref. [2], please use GPUMD-v2.6.

References:

• [1] Zheyong Fan, Improving the accuracy of the neuroevolution machine learning potentials for multi-component systems, To be submitted.
• [2] Zheyong Fan, Zezhu Zeng, Cunzhi Zhang, Yanzhou Wang, Keke Song, Haikuan Dong, Yue Chen, and Tapio Ala-Nissila, Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport, Phys. Rev. B. 104, 104309 (2021). https://doi.org/10.1103/PhysRevB.104.104309

GPUMD-v2.6

Major changes:

• Added the Neuroevolution machine learning potentials (NEP) [https://arxiv.org/abs/2107.08119].
• Fixed a bug about the simulation cell. The volume of some simulation cells were calculated to be negative, which can give rise to a "negative thermal conductivity". Otherwise, the dynamics is not affected.
• Future work: we will release GPUMD-v2.7 soon, where the accuracy of the NEP potential for multi-component systems will be significantly improved.

GPUMD-v2.5.1

Major changes compared to GPUMD-v2.5:

• Created Jupyter notebooks (with Python 3) as the new tutorials. The thermo package is used.
• Fixed a bug related to the ensemble npt_ber command. The bug will affect simulations where the z direction is periodic and not deformed and at the same time the target pressures for the z and x directions are different. The reason is that the unfix code will control the box size in the z direction according to the target pressure in the x direction.

GPUMD-v2.5

Major changes compared to GPUMD-v2.4.1:

• Added the Tersoff-mini potential.
• Added the FCP potential.
• Added two modal analysis methods (GKMA and HNEMA).
• Changed the data format in the xyz.in file by removing the has_layer item in the first line.
• Changed the grammar of the potential keyword and removed the potentials keyword.
• Added the minimize keyword.
• Changed the grammar of the compute_shc keyword.
• Enriched the features of the compute_dos keyword.
• Added support for dumping the netCDF file.
• Added support for Windows with MSVC.

GPUMD-v2.4.1

The file makefile.phonon in GPUMD-v2.4 is incorrect.
It has been corrected in GPUMD-v2.4.1.

GPUMD-v2.4

Major changes:

1. Added a main program for phonon calculations based on harmonic lattice dynamics.
2. Improved the calculation of the vibrational density of states from the velocity autocorrelation function.

GPUMD-v2.3

1. Added the general Tersoff potential (called Tersoff1988), which is applicable to systems with an arbitrary number of atom types. The added version is as general as the Tersoff potential in LAMMPS.
2. Added the support of triclinic box (but it currently cannot be used together with the NPT ensemble).

GPUMD-v2.2

1. Improved the code structure and fixed a few memory bugs introduced in versions 2.0 and 2.1.
2. Changed the data format in the xyz.in file and removed the layer.in file.
3. Changed the output of the dump_position command from xyz.out to move.xyz.
4. Added a dump_restart command which can produce a restart.out file.
5. The "compute_temp" command has been replaced by the more general "compute" command.
6. Added the "deform" command, which can be used to compute the stress-strain relation.