Compiler Optimization

Compiler optimization aims to automatically improve the performance and efficiency of compiled code, focusing on speed, memory usage, and code size. Current research heavily utilizes machine learning, particularly large language models and Bayesian optimization, to automate the complex process of selecting optimal compiler settings and transformations, often surpassing the capabilities of hand-crafted optimization strategies. This field is crucial for advancing software performance across various domains, from general-purpose computing to specialized applications like deep learning and quantum computing, enabling faster and more efficient software execution.

Papers