Tensor Processing Primitives: A Programming Abstraction for Efficiency and Portability in Deep Learning and HPC Workloads.
Evangelos GeorganasDhiraj D. KalamkarSasikanth AvanchaMenachem AdelmanDeepti AggarwalCristina AndersonAlexander BreuerJeremy BruestleNarendra ChaudharyAbhisek KunduDenise KutnickFrank LaubVasimuddin MdSanchit MisraRamanarayan MohantyHans PabstBrian RetfordBarukh ZivAlexander HeineckePublished in: Frontiers Appl. Math. Stat. (2022)