OMP2HMPP: Compiler Framework for Energy Performance Trade-off Analysis of Automatically Generated Codes

TL;DR

OMP2HMPP is a compiler framework that automatically transforms OpenMP code into various HMPP versions, evaluates their performance and power consumption, and helps select the most efficient code variant, achieving significant speed-ups and energy savings.

Contribution

It introduces an automated method for exploring energy-performance trade-offs in code transformations from OpenMP to HMPP, matching expert manual optimizations.

Findings

Achieves an average speed-up of 31x over original code.

Attains a 5.86x improvement in operations per watt.

Automatically finds transformations equivalent to expert manual tuning.

Abstract

We present OMP2HMPP, a tool that, in a first step, automatically translates OpenMP code into various possible transformations of HMPP. In a second step OMP2HMPP executes all variants to obtain the performance and power consumption of each transformation. The resulting trade-off can be used to choose the more convenient version. After running the tool on a set of codes from the Polybench benchmark we show that the best automatic transformation is equivalent to a manual one done by an expert. Compared with original OpenMP code running in 2 quad-core processors we obtain an average speed-up of 31x and 5.86x factor in operations per watt.