Loo.py: From Fortran to performance via transformation and substitution rules

TL;DR

Loo.py is a transformation system that enables high-performance, portable GPU programming by allowing user-controlled transformations of array programs, including legacy Fortran code, through Python-based rules.

Contribution

It introduces a flexible transformation framework that extends to legacy languages like Fortran, facilitating performance optimization on data-parallel architectures.

Findings

Supports transformation of Fortran and other languages

Enables high-performance GPU code generation

Offers a flexible, user-controlled transformation mechanism

Abstract

A large amount of numerically-oriented code is written and is being written in legacy languages. Much of this code could, in principle, make good use of data-parallel throughput-oriented computer architectures. Loo.py, a transformation-based programming system targeted at GPUs and general data-parallel architectures, provides a mechanism for user-controlled transformation of array programs. This transformation capability is designed to not just apply to programs written specifically for Loo.py, but also those imported from other languages such as Fortran. It eases the trade-off between achieving high performance, portability, and programmability by allowing the user to apply a large and growing family of transformations to an input program. These transformations are expressed in and used from Python and may be applied from a variety of settings, including a pragma-like manner from other languages.