Automatic feature generation for machine learning--based optimising compilation

Hugh Leather, Edwin V. Bonilla, Michael O’Boyle

ACM Transactions on Architecture and Code Optimization 2014 article cited by 31

Source Open-access PDF More in Software & Coding

importance 4/10 ai assisted Parallel Computing and Optimization Techniques compiler optimization stage: research

Problem

Machine-learned compiler optimisation heuristics depend on hand-chosen program features, and the space of possible features is effectively infinite.

Why it matters

Manual feature selection bottlenecks the automation of compiler optimisation, affecting performance of all compiled software.

What the AI did

A grammar-described feature space is searched with genetic programming and predictive modelling to automatically find features that improve the learned heuristic, applied to loop unrolling in GCC 4.3.1.

Results

not stated in source

Impact

Applied to loop unrolling in GCC 4.3.1 on a Pentium 6, but measured outcomes are cut off from the abstract.

Record

Venue	ACM Transactions on Architecture and Code Optimization
Published	2014
Work type	article
Authors	Hugh Leather, Edwin V. Bonilla, Michael O’Boyle
Citations	31 (OpenAlex, at import time)
References	27
Model or method	genetic programming with predictive modelling
Organization	not stated in source
Primary source	https://doi.org/10.1145/2536688
Open access	bronze
Keywords	compiler optimisation, feature generation, genetic programming, machine learning, GCC, loop unrolling

Importance rating

4 / 10

real but narrower advance

Feature selection bottlenecks machine-learned compiler optimisation affecting compiled software performance; abstract quantifies gains (76% vs 59% of max speedup on 57 programs) but claims no industry-wide scale.

Basis: stated. The source itself states the scale. Stated scale: not stated in source. Method and limits.

More in Software & Coding

#	Paper	Year	Citations	Importance
1	ImageNet classification with deep convolutional neural networks Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton	2017	106,345	8
2	Deep Residual Learning for Image Recognition Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun	2016	221,134	8
3	Faster sorting algorithms discovered using deep reinforcement learning Daniel J. Mankowitz, Andrea Michi, Anton Zhernov, Marco Gelmi, Marco S...	2023	175	8
4	A deep learning‐based approach for software vulnerability detection using code metrics Fazli Subhan, Xiaoxue Wu, Lili Bo, Xiaobing Sun, Muhammad Rahman	2022	42	7
5	Deep Learning-Based Attack Detection and Classification in Android Devices Alfonso Gómez, Antonio Muñoz	2023	41	7