2026
QiMeng-Kernel: Macro-Thinking Micro-Coding Paradigm for LLM-Based High-Performance GPU Kernel Generation
AAAI 2026technical
Developing high-performance GPU kernels is critical for AI and scientific computing, but remains challenging due to its reliance on expert crafting and poor portability. While large language models (LLMs) offer promise for automation, both general-purpose and finetuned LLMs suffer from two fundament