Several minor changes in the CG eigensolver

Description

Several minor changes in the CG eigensolver:

• Reduce the memory used by the CG eigensolver.
• Improve speed when the fold option is passed, by working with arrays of size np and not np_part
• Removing unnecessary initialization and zero-filing.
• Using Blas when possible This makes the CG part few percent faster (~4%) for test systems 1ALA, 2ALA, and 3ALA.

News snippet

Minor improvement of the CG eigensolver.

Checklist

• I have checked that my code follows the Octopus coding standards
• I have added tests for all the new features added in this request.