Time skewing made simple

Abstract

Time skewing and loop tiling has been known for a long time to be a highly \nbeneficial acceleration technique for nested loops especially on bandwidth \nhungry multi-core processors, but it is little used in practice because \nefficient implementations utilize complicated code and simple or abstract ones \nshow much smaller gains over naive nested loops. We break this dilemma with an \nessential time skewing scheme that is both compact and fast.