ECO: Enhanced Code Optimization via Performance-Aware Prompting for Code-LLMs

2025-10-12

AI Summary

Existing slow-fast code-pair-based methods fail to uncover the causal mechanisms underlying performance improvements, causing models to merely mimic superficial patterns without developing deep performance reasoning capabilities. To address this, we propose a fine-tuning-free, plug-and-play performance-aware prompting framework. First, a Root-cause Oriented Insight (ROI) mechanism identifies optimization root causes and underlying reasoning logic. Then, a symbolic advisor module—augmented with retrieval-enhanced context—is integrated to generate customized prompts for bottleneck diagnosis, differential analysis, and instruction extraction. This framework explicitly guides large language models to model performance drivers, enabling intelligent rewriting from inefficient to high-performance code. Experiments demonstrate up to 7.81× runtime speedup across multiple tasks while strictly preserving functional correctness.

Technology Category

• Search and Optimization: Metareasoning and Metaheuristics
• Computer Vision: Visual Reasoning & Symbolic Representations
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Code runtime optimization-the task of rewriting a given code to a faster one-remains challenging, as it requires reasoning about performance trade-offs involving algorithmic and structural choices. Recent approaches employ code-LLMs with slow-fast code pairs provided as optimization guidance, but such pair-based methods obscure the causal factors of performance gains and often lead to superficial pattern imitation rather than genuine performance reasoning. We introduce ECO, a performance-aware prompting framework for code optimization. ECO first distills runtime optimization instructions (ROIs) from reference slow-fast code pairs; Each ROI describes root causes of inefficiency and the rationales that drive performance improvements. For a given input code, ECO in parallel employs (i) a symbolic advisor to produce a bottleneck diagnosis tailored to the code, and (ii) an ROI retriever to return related ROIs. These two outputs are then composed into a performance-aware prompt, providing actionable guidance for code-LLMs. ECO's prompts are model-agnostic, require no fine-tuning, and can be easily prepended to any code-LLM prompt. Our empirical studies highlight that ECO prompting significantly improves code-LLMs' ability to generate efficient code, achieving speedups of up to 7.81x while minimizing correctness loss.

Problem

• Addressing runtime optimization challenges requiring performance trade-off reasoning
• Overcoming superficial pattern imitation in code optimization methods
• Enhancing code-LLMs' ability to generate genuinely efficient code

Innovation

• Distills runtime optimization instructions from code pairs
• Uses symbolic advisor for bottleneck diagnosis
• Composes performance-aware prompts for code-LLMs