PIE

The PIE benchmark evaluates code performance optimization by measuring the effectiveness of various models in enhancing runtime speed and optimization rates based on a curated collection of real-world optimization trajectories.

Papers using PIE (3)

MaxCode: A Max-Reward Reinforcement Learning Framework for Automated Code Optimization2026PerfCoder: Large Language Models for Interpretable Code Performance Optimization2025PerfRL: A Small Language Model Framework for Efficient Code Optimization2023 · 2 cites