abacus crushed with HSE band calculation

[zhang-changwei]

Describe the bug

Black phosphorus primitive cell. HSE+DFT-D3(0) functional. I'm trying to run a scf cacluation with a gamma centered kmesh and several extra zero weight kpoints, just like what should be done in vasp. However, abacus crushes even I only use one mpi process. If I change the functional to PBE, abacus seems work well. Part of the output is showed as follows:

 -------------------------------------------
 SELF-CONSISTENT : 
 -------------------------------------------
 START CHARGE      : atomic
abacus: /public1/home/sch3658/src/ABACUS/abacus-develop_deepmodeling-develop/deps/LibRI/include/RI/physics/../distribute/Split_Processes.hpp:31: std::tuple<int, unsigned long> RI::Split_Processes::split(const int &, const unsigned long &): Assertion `rank_size>=group_size' failed.

The full input files are contained in the attachment. BP.zip

Expected behavior

No response

To Reproduce

No response

Environment

abacus 3.2.0

Additional Context

No response

[mohanchen]

For further discussions, we will move to https://github.com/deepmodeling/abacus-develop