OpenMPI 5.0.3 in gcc-toolchain cannot do OMP parallel calculation normally

[QuantumMisaka]

Describe the bug

Now the gcc toolchain in abacus repo is using OpenMPI 5.0.3. In my test, this version of OpenMPI and coressponding toolchain cannot do OMP parallel calculation normally. if OMP_NUM_THREADS>1, the SCF comsusing time in each eletronic step will be very large. for example:

Time	MPI=16, OMP=4	MPI=64, OMP=1
2H-Au111	384 s	5.4 s

Expected behavior

ABACUS by gcc-toolchain can use OpenMP parallel part (at least OMP_NUM_THREADS=4 in OpenMPi 4.1.5)

To Reproduce

gcc toolchain with gcc 12.2.0

Environment

No response

Additional Context

The easiest way is to downgrade the version of OpenMPI, but is there another way ?

Task list for Issue attackers (only for developers)

• Verify the issue is not a duplicate.
• Describe the bug.
• Steps to reproduce.
• Expected behavior.
• Error message.
• Environment details.
• Additional context.
• Assign a priority level (low, medium, high, urgent).
• Assign the issue to a team member.
• Label the issue with relevant tags.
• Identify possible related issues.
• Create a unit test or automated test to reproduce the bug (if applicable).
• Fix the bug.
• Test the fix.
• Update documentation (if necessary).
• Close the issue and inform the reporter (if applicable).

[QuantumMisaka]

OpenMPI 5.0.5 solved this problem
Update OpenMPI to 5.0.5 in PR #5088

[QuantumMisaka]

Supplement: OpenMPI 5.0.5 do not solve this problem actually
but I've no idea how to solve it.