ABACUS FIRE is much inefficient by ASE FIRE

[QuantumMisaka]

Details

As #4016 , ABACUS FIRE algorism need some improvements

One example as follow, do FIRE optization task from a same STRU with same INPUT setting (DFT is using R2SCAN in LibXC), while ASE-ABACUS FIRE using 49 ION step but ABACUS FIRE cannot converge in 300 steps with largest gradient stuck around 0.3 eV/Ang

FIRE-SCAN-ase.tar.gz

ABACUS-SCAN-FIRE.tar.gz

Task list for Issue attackers (only for developers)

• [ ] Reproduce the performance issue on a similar system or environment.
• [ ] Identify the specific section of the code causing the performance issue.
• [ ] Investigate the issue and determine the root cause.
• [ ] Research best practices and potential solutions for the identified performance issue.
• [ ] Implement the chosen solution to address the performance issue.
• [ ] Test the implemented solution to ensure it improves performance without introducing new issues.
• [ ] Optimize the solution if necessary, considering trade-offs between performance and other factors (e.g., code complexity, readability, maintainability).
• [ ] Review and incorporate any relevant feedback from users or developers.
• [ ] Merge the improved solution into the main codebase and notify the issue reporter.

[WHUweiqingzhou]

@YuLiu98 could you have a look?

[YuLiu98]

The original FIRE is not implemented by me. I just reconstructed it during MD reconstruction. So I'm not familiar with the algorithm.