Unable to perform the all-electron calculations

Dear MRCC developers,
Please see the attached output: the keyword "core" was set to 0, parsed correctly; however, the "Number of core electrons" remains 2. This problem holds in other cases, too.
Latest 25.1.1 binary.

UPD. I've noticed that the energies of core 0 and core frozen are actually different. So, it's just an output bug?

Yes, the number of core electrons is 2. But they are not correlated. I cannot see any problem either with the output or with the calculation.

kallay wrote:

Yes, the number of core electrons is 2. But they are not correlated. I cannot see any problem either with the output or with the calculation.
 

What is the difference between core equal 0 and core equal frozen (attachement) then? The energy from the former calculation is significantly lower than from the latter, like the core electrons are correlated in the former case and frozen in the latter.
(It seems that forum does not allow equal sign.)

Please see the description of keyword core in the manual.

Let me explain what causes some confusion in the program output in more detail.
Consider a larger system such as benzene. The frozen core is standardly defined as six doubly occupied 1s orbitals of carbons, i.e. 12 electrons.

- We can perform a correlated calculation with core eq frozen. The program prints “Number of core electrons:        12”, and the calculation converges to some energy of -231.581 Ha.

- Well, we can perform the calculation with core eq 6. Expectedly, this will result in six orbitals being dropped, so the same calculation as above applies. The program prints “Number of core electrons:        12”, and the calculation converges to the same energy of -231.581 Ha.

- Let us include some subvalence correlation: core eq 4. Expectedly, the program prints “Number of core electrons:        8”, and the calculation converges to a lower energy of -231.585 Ha.

- Moving forward: core eq 1. Everything is still ok. The program prints “Number of core electrons:        2”, and the calculation converges to an even lower energy of -231.593 Ha.

- Finally, we can correlate all electrons in this system: core eq 0. What could one expect in the output, considering the above calculations? I expect “Number of core electrons:        0”. However, the program prints “Number of core electrons:        12”, but I guess, actually performs what was requested, and the calculation converges to the lowest energy of -231.596 Ha. This is, however, not obvious because the number of correlated electrons is not printed out explicitly somewhere else. One can assume that the number of core electrons was again defaulted to 12 for some reason (e.g. some combination of keywords overwrites corr).

 

The "Number of core electrons:" is not necessarily the same as the number of frozen core orbitals used in the correlation calculation.

You are right, the core=0 correlates all electrons. You can see the number of correlated orbitals in the output of the ccsd program:
Number of occupied/virtual orbitals:

Additional suggestion: you are using the older, 4-center CCSD(T) version which is far less optimized. I would strongly recommend the highly-optimized DF-based code available via:
calc=DF-CCSD(T)