Comprehensive Optimization of Parametric Kernels for Graphics Processing Units

TL;DR

This paper presents a method to optimize GPU programs by treating hardware and program parameters as unknowns during code generation, aiming to enhance robustness, portability, and efficiency through algebraic techniques.

Contribution

It introduces a novel approach that uses symbolic parameters and case-based optimization to improve GPU program performance and adaptability.

Findings

Preliminary experiments show promising results

Optimization improves program robustness and portability

Method leverages recent advances in computer algebra

Abstract

This work deals with the optimization of computer programs targeting Graphics Processing Units (GPUs). The goal is to lift, from programmers to optimizing compilers, the heavy burden of determining program details that are dependent on the hardware characteristics. The expected benefit is to improve robustness, portability and efficiency of the generated computer programs. We address these requirements by: (1) treating machine and program parameters as unknown symbols during code generation, and (2) generating optimized programs in the form of a case discussion, based on the possible values of the machine and program parameters. By taking advantage of recent advances in the area of computer algebra, preliminary experimentation yield promising results.