Omniwise: Predicting GPU Kernels Performance with LLMs
Published in arXiv, 2025
This work explores using large language models to predict the performance characteristics of GPU kernels, enabling more intelligent optimization decisions.
Recommended citation: Z. Wang, C. Ramos, M. A. Awad, K. Lowery. "Omniwise: Predicting GPU Kernels Performance with LLMs." arXiv preprint arXiv:2506.20886.
Download Paper