wqzhou

Bugs: the results of different parallel schemes vary greatly for LCAO calculations

3

### Describe the bug During the test of issue #4058, I find results of different parallel settings are totally different for same INPUT: ```bash OMP_NUM_THREADS=1 mpirun -np 16 abacus |...

Bugs: Different SCF with equivalent `STRU`

8

### Describe the bug In Issue #3954, the user find with STRU1 and STRU2, ABACUS give different SCF. One of them cannot be converged: STRU1: ```bash Ni 1.0000000000 48 0.0000000000...

### Background After refactor, current `print_psi()` is useful only for `out_wfc_flag=1`. However it still occpies some memory and time cost (not much) in `wfc_2d_to_grid` calculations. Maybe it can be refactored,...

Feature: unify the implementation of charge mixing in SCF and mixing used in EXX

1

### Background Currently, the mixing used in EXX is independent with the charge mixing in SCF. Maybe a unified mixing is helpful for better convergence of HSE calculations. ### Describe...

Test: should we also make CI test by using Intel Env?

6

### Describe the Testing Issue Now we only have CI tests under GNU Env. ### Additional Context _No response_ ### Task list for Issue attackers (only for developers) - [...

Fix #4802. ### List of Changes 1. replace replace `dsytrf` & `dsytri` & `dgemv` by `dsysv` in Broyden mixing. Plz note that The pass of PR #4842 CI tests show...

For test

### Background For SOC accuracy tests, this feature must be supported. ### Describe the solution you'd like a feature to read charge density and rotate easy-axis for SOC nscf calcualtions...

### Describe the Testing Issue After #4493, warnings are not allowed anymore in CI tests. However, #4493 only changed threshold to make CI pass, instead of changing refs themselves. At...

Feature: update LibComm to the latest version

3

### Background Using latest version of LibComm can fix #4310 and #3674. ### Describe the solution you'd like update LibComm ### Task list only for developers - [ ] Notice...