II-NEW: Research Software Infrastructure for Tensor Computations

点击翻译按钮获取中文摘要

The goal of this project is to build a new software infrastructure forcontinuing research on compiler optimizations for tensor computations.Motivated by the successes of the model-driven search-basedoptimization approach of the Tensor Contraction Engine (TCE) and thepolyhedral model-based transformations, an optimization infrastructureis being developed that combines the key aspects of the TCE and thepolyhedral models and provides the flexibility to continue research onoptimizing and automatically parallelizing tensor computations forparallel and/or distributed computations for any machine architecture,including multi-cores and GPUs.This novel optimization framework has potential for high payoffs inenabling future research and eventually in generating high-performancecode for an important class of quantum chemistry codes and othertensor computations. After further research, it will become avaluable tool for reducing the time to develop high-performanceapplications in several areas of science and engineering.