I am doing calculation on a bulk periodic system using VASP, but do not know what is the reference of those absolute values for conduction band minimum and valence band maximum.
Use VASPKIT to know the HOMO/LUMO and bandgap. There are also several python scripts that can obtain bandgap as well as CBM and VBM values. Search for Vasp python scripts for post-processing in Github.
Pritam Kumar Panda , I do know the band gap as well as CBm and VBM after doing the VASP calculations, what I was concern about is the reference of those CBm and VBM. We only get the absolute values of VBM and CBm after running any script for post-processing, but I do not know from which reference those values are obtained. For example, we can know the CBm w.r.t the vacuum after doing the slab calculations. However, in case of bulk can there exist any such kind of reference without doing the slab calculations?
I admittedly do not understand your question: since this is a bulk calculation there is not reference, that is the only important thing is the location of the Fermi Energy and the VBM-CBM positions w.r.t. that. This gives you the character of the material (conducting, semiconducting, insulating) and you can work form there.
Indeed, different pseudo potentials (let's say to different PBE) might give different absolute values of the CBM, VBM and Fermi energy, but their relative positions (i.e. the band gap) should not change (much) for a well converged calculations.
In bulk system, VASP assigns an arbitrary zero energy reference level, so the absolute values of VBM and CBM do not carry any meaning. This means that for different systems, you cannot compare their VBM/CBM position directly. In case of curiosity, VASP set the zero energy reference level in a way that the average potential from all atoms in the unit cell is zero, i.e., the Fourier component for G=0 is zero. If you understand, then you may realize that different systems have different average potential so that different systems have different zero energy reference level, and that's the reason why I say "VASP assigns an arbitrary zero energy reference level".