GPUMD-v3.9.5

[brucefan1983]

• Deprecated:

  • [GPUMD and NEP] NEP2 is not supported any more. Now there are only NEP3 and NEP4. No nep2 #689 totally remove NEP2 #687

• New or enhanced features:

  • [NEP] Introduced use_typewise_cutoff and use_typewise_cutoff_zbl keywords for NEP training, which could lead to more flexible cutoff schemes and improved performance. typewise cutoffs #660 typewise cutoff factors #684
  • [GPUMD] Allow for creating multiple xyz files using dump_exyz during a run. Add separated parameter for dump_exyz #629
  • [GPUMD] Output the total electron stopping power loss during the run with electron_stop. calculate stopping_power_loss #634
  • [GPUMD] Add electric field (add_efield) based on given fixed charges. Add efield #662
  • [GPUMD] Add force (add_force) to a group of atoms. Add force #658
  • [GPUMD] When velocity is missing from run.in, now the code initializes the velocities with a temperature of 300 K, instead of showing the previous "undefined" behavior.
  • [GPUMD] correct_velocity can now be used to correct the velocities for different groups of atoms. correct velocity for different groups #668
  • Better error reporting when reading training/test datasets. report error from train.xyz #699 Check nan and inf #697

• Bug fixes:

  • [GPUMD] For the compute_rdf command, stop calculation when any periodic thickness is smaller than 2.5 times of the RDF cutoff. Fixed bug in RDF calculations using small boxes #649
  • [GPUMD] Fix a bug introduced after adding dipole/polarizability prediction in GPUMD-v3.9.2. The NEP energy was wrongly calculated when multi-GPU is used for MD runs. fix energy bug for multi-gpu #661

---

This discussion was created from the release GPUMD-v3.9.5.