enh: added parallel-k calculations for NC/SOC
This should really speed up calculations with NC/SOC since many systems are small in size and will benefit from the dense LAPACK calls rather than lots of communications.
I have added more tests to check that the implementations yield the same results.
I have also reduced memory requirements since the auxilliary arrays were never used in the size they were supplied.
Also fixed the SOC write-out of CITATIONS.bib, now it correctly writes to the slabel.bib
Fixes #33 (closed).