Making LLMs Optimize Multi-Scenario CUDA Kernels Like Experts

TL;DR

This paper introduces CUDAMaster, a system that automates GPU kernel optimization across multiple scenarios, significantly improving performance and surpassing existing methods like Astra and cuBLAS.

Contribution

It develops MSKernelBench, a comprehensive benchmark for multi-scenario GPU kernels, and proposes CUDAMaster, a hardware-aware, multi-agent system for automated kernel optimization.

Findings

CUDAMaster achieves about 35% speedup over Astra.

It matches or exceeds the performance of cuBLAS in several cases.

The benchmark covers diverse applications including scientific computing routines.

Abstract

Optimizing GPU kernels manually is a challenging and time-consuming task. With the rapid development of LLMs, automated GPU kernel optimization is gradually becoming a tangible reality. However, current LLM-driven automated optimization methods narrowly focus on machine learning applications, such as PyTorch operator optimization, while overlooking broader domains like sparse matrix operations in scientific computing. Extending to these broader applications brings new challenges for the benchmark and algorithm. Therefore, developing a general-purpose automated kernel optimization method becomes our primary focus. In this paper, we address the absence of systematic evaluation for multi-scenario settings by introducing MSKernelBench, which spans multiple scenarios, including fundamental algebraic operations, common LLM kernels, sparse matrix operators, and scientific computing routines, each supporting both FP32 and BF16 precision. Building on this benchmark, we introduce CUDAMaster, a multi-agent, hardware-aware system for kernel optimization that leverages profiling information and automatically constructs the full compilation and execution toolchain. Experimental results demonstrate that CUDAMaster achieves significant speedups across most operators, outperforming Astra by about 35%. In several cases, its performance matches or surpasses that of highly optimized, closed-source libraries such as cuBLAS. A demo showcasing the original and optimized code for each operator is available at https://hanyx2021.github.io/MSKernelBenchDemo/.