[Bug] `Eigenvalues are too large` during calculation of SiC with `dav_subspace`

[ZhouXY-PKU]

Describe the bug

SiC with efield and nspin=2.

At RELAX STEP 2, while EDIFF is 1e-5 eV; DRHO is 1e-7; about to converge.
ITER TMAG AMAG ETOT/eV EDIFF/eV DRHO TIME/s

Expected behavior

No response

To Reproduce

ABACUS-3.11.0-beta8
On GPU A800*2

SiC_bug.tar.gz

Environment

No response

Additional Context

No response

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).

[mohanchen]

It's strange. More investigations are needed.