Convergence difficulty met in LiCoO calculation with the normal nupdown setting

[hongriTianqi]

Describe the bug

Convergence difficulty met in LiCoO calculation with the nupdown setting.
The test is presented here:
https://app.bohrium.dp.tech/abacustest?request=GET%3A%2Fapplications%2Fabacustest%2Fjobs%2Fjob-abacustest-v0.3.95-eb7918

with 34/100 cases not converged.

Expected behavior

All pass.

To Reproduce

No response

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).

[WHUweiqingzhou]

I think this bad convergence comes from mixing_restart=10, which is a very bad initial guess.
I believe this issue can be solved after PR #3681 by setting mixing_restart = 1e-3.

[WHUweiqingzhou]

I am still working on this issue.

[rimujiangshu]

I can not open the link，so maybe you can resubmit the issue again.

[mohanchen]

Since we cannot get the data (don't believe any link beyond github that can save the data in a considerable amount of time), I will transfer this issue to discussion