TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators

TL;DR

TritonBench is a new benchmark designed to evaluate large language models' ability to generate efficient Triton GPU kernels, revealing current models' limitations in high-performance code generation.

Contribution

This work introduces TritonBench, the first comprehensive benchmark for Triton operator generation, combining correctness and efficiency evaluation on real-world and PyTorch-aligned operators.

Findings

Current LLMs struggle to generate efficient Triton code.

TritonBench provides a systematic evaluation framework.

Significant gap identified in high-performance Triton code generation.

Abstract

Triton, a high-level Python-like language designed for building efficient GPU kernels, is widely adopted in deep learning frameworks due to its portability, flexibility, and accessibility. However, programming and parallel optimization still require considerable trial and error from Triton developers. Despite advances in large language models (LLMs) for conventional code generation, these models struggle to generate accurate, performance-optimized Triton code, as they lack awareness of its specifications and the complexities of GPU programming. More critically, there is an urgent need for systematic evaluations tailored to Triton. In this work, we introduce TritonBench, the first comprehensive benchmark for Triton operator generation. TritonBench features two evaluation channels: a curated set of 184 real-world operators from GitHub and a collection of operators aligned with PyTorch interfaces. Unlike conventional code benchmarks prioritizing functional correctness, TritonBench also profiles efficiency performance on widely deployed GPUs aligned with industry applications. Our study reveals that current state-of-the-art code LLMs struggle to generate efficient Triton operators, highlighting a significant gap in high-performance code generation. TritonBench will be available at https://github.com/thunlp/TritonBench.