Copious generation of electron-positron pair plasma is probably the
most crucial part of (almost) all pulsar models - it is widely agreed
upon than a neutron star is active as radio pulsar when pairs are
produced in polar cap cascades and ceases when cascades die.  In this
talk I will report on modeling of physical processes in pulsar polar
caps based on the most recent self-consistent numerical models of
particle acceleration.  It turned out that particles are accelerated
more efficiently, can produce more pairs, and the escaping
gamma-radiation is in lower energy bands than it was predicted by
previous models; there is however an upper limit on the multiplicity
of pair plasma.  I will discuss implication of these results for
pulsar gamma-ray emission and for phenomena depending on pulsar pair
yield, such as pulsar wind nebulae and positron content of cosmic
rays.