Mass die-off or steady culling

A (probably naive) question for my science/math/software-oriented friends: In genetic algorithms, which is better: mass die-offs or steady culling? Or are they mathematically equivalent?