For really big quantities of pure CCC plasmid DNA, I guess CsCl gradient centrifugation is still the method of choice, given that you have access to an ultra-centrifuge.
Cesium chloride + Ethidium bromide density gradient centrifugation. Ethidium bromide is essential to the procedure as linear/nicked DNA will intercalate a greater amount of the ethidium, than circular/supercoiled DNA. This causes the linear DNA to have a lower density than the closed circular DNA and allows the separation that occurs in the density gradient.