Towards a learning-based performance modeling for accelerating Deep Neural Networks

TL;DR

This paper explores machine learning-based performance models to optimize CNN computations, demonstrating that predictive models can outperform manually optimized convolution operators on ARM Mali GPUs.

Contribution

It introduces a machine learning approach to performance modeling for CNNs, specifically applying decision trees and Bayesian classifiers to improve operator selection.

Findings

Predictive models outperform manual operator selection.

Models built using decision trees and Bayesian classifiers.

Validation on ARM Mali GPU shows improved performance.

Abstract

Emerging applications such as Deep Learning are often data-driven, thus traditional approaches based on auto-tuners are not performance effective across the wide range of inputs used in practice. In the present paper, we start an investigation of predictive models based on machine learning techniques in order to optimize Convolution Neural Networks (CNNs). As a use-case, we focus on the ARM Compute Library which provides three different implementations of the convolution operator at different numeric precision. Starting from a collation of benchmarks, we build and validate models learned by Decision Tree and naive Bayesian classifier. Preliminary experiments on Midgard-based ARM Mali GPU show that our predictive model outperforms all the convolution operators manually selected by the library.