This repository contains tools and libraries for building AIR platforms, runtimes and compilers.

• Building
• Running Examples
• AIRCC

• AIR Dialect
• AIRRt Dialect
• AIR Transform Passes
• AIR Conversion Passes
• Transform Dialect Extensions

• AIR Compute Model — formal semantics of the AIR hierarchy and memory model, with NPU and GPU backend mappings

• AIR Asynchronous Concurrency
• AIR Runner

If you use MLIR-AIR in your research, please cite our paper:

E. Wang, S. Bayliss, A. Bisca, Z. Blair, S. Chowdhary, K. Denolf, J. Fifield, B. Freiberger, E. Hunhoff, P. James-Roxby, J. Lo, J. Melber, S. Neuendorffer, E. Richter, A. Rosti, J. Setoain, G. Singh, E. Taka, P. Vasireddy, Z. Yu, N. Zhang, J. Zhuang. "From Loop Nests to Silicon: Mapping AI Workloads onto AMD NPUs with MLIR-AIR". arXiv:2510.14871, October 2025.

Copyright© 2019-2023 Advanced Micro Devices, Inc.