Evaluating Language Models for Efficient Code Generation

TL;DR

This paper introduces Differential Performance Evaluation (DPE), a new framework for reliably assessing the efficiency of Large Language Models in code generation, addressing limitations of existing benchmarks.

Contribution

The paper presents DPE, a novel evaluation framework that uses efficiency-demanding tasks and compound metrics to better measure LLM code efficiency, along with the EvalPerf benchmark.

Findings

Model size has limited impact on code efficiency.

Instruction tuning improves both correctness and efficiency.

EvalPerf is a reliable and platform-independent benchmark.

Abstract

We introduce Differential Performance Evaluation (DPE), a framework designed to reliably evaluate Large Language Models (LLMs) for efficient code generation. Traditional coding benchmarks often fail to provide reliable insights into code efficiency, due to their reliance on simplistic test inputs and the absence of effective compound metrics. DPE addresses these issues by focusing on efficiency-demanding programming tasks and establishing an insightful compound metric for performance evaluation. DPE operates in two phases: To curate efficiency datasets, it selects efficiency-demanding tasks from existing coding benchmarks and generates computationally expensive inputs to stress the efficiency of LLM solutions. To assess the code efficiency, DPE profiles the new solution and compares it globally against a set of reference solutions that exhibit distinct efficiency levels, where the matched level defines its efficiency score. As a proof of concept, we use DPE to create EvalPerf, a benchmark with 121 performance-challenging coding tasks. Our comprehensive evaluation draws interesting findings on the efficiency impact of model sizes, instruction tuning, and prompting. For example, while the scaling law fails to account for code efficiency, general instruction tuning benefits both code correctness and efficiency. We also evaluate the evaluation by examining the effectiveness of DPE, showing that EvalPerf is reliable and convenient to use even across platforms.