ML$^2$Tuner: Efficient Code Tuning via Multi-Level Machine Learning Models

TL;DR

ML$^2$Tuner is a novel multi-level machine learning approach that significantly improves deep learning model autotuning efficiency by predicting valid configurations and accurately estimating performance, reducing tuning time and invalid attempts.

Contribution

It introduces a multi-level ML tuning framework with validity and performance prediction models, enhancing autotuning efficiency for deep learning accelerators.

Findings

Achieves similar performance with only 12.3% of samples compared to TVM.

Reduces invalid profiling attempts by an average of 60.8%.

Demonstrates effectiveness on an extended VTA accelerator.

Abstract

The increasing complexity of deep learning models necessitates specialized hardware and software optimizations, particularly for deep learning accelerators. Existing autotuning methods often suffer from prolonged tuning times due to profiling invalid configurations, which can cause runtime errors. We introduce ML$^2$Tuner, a multi-level machine learning tuning technique that enhances autotuning efficiency by incorporating a validity prediction model to filter out invalid configurations and an advanced performance prediction model utilizing hidden features from the compilation process. Experimental results on an extended VTA accelerator demonstrate that ML$^2$Tuner achieves equivalent performance improvements using only 12.3% of the samples required with a similar approach as TVM and reduces invalid profiling attempts by an average of 60.8%, Highlighting its potential to enhance autotuning performance by filtering out invalid configurations