基于 MLIR 的开源编译器工具链，提供贴近硬件的 Python API，用来给 AMD Ryzen AI NPU（AI Engine）编写和编译算子程序。
面向手机端的高性能神经网络推理框架，针对 ARM 平台深度优化，让 Caffe、ONNX 等模型能在 Android/iOS 上低延迟运行。
Triton 是一门面向 GPU 编程的类 Python 语言及其编译器，让开发者不用写 CUDA 就能写出接近手写性能的高效内核。
基于 MLIR 的可重定向机器学习编译器与运行时工具包，把 PyTorch、TensorFlow、JAX、ONNX 等模型编译到 CPU、CUDA、ROCm、Vulkan 等多种硬件上运行。
字节跳动开源的模型编译方案，基于 MLIR/LLVM 把 PyTorch、ONNX、TensorFlow 等模型编译到多种硬件上运行，解决跨平台部署和推理性能问题。
收录 CUDA、cuBLAS、CUTLASS、TensorRT-LLM、Triton、TVM、MLIR 等 GPU 加速与高性能计算开源项目的精选清单，方便按方向找底层算子、编译器和推理框架。
…ler measurements and added support for preserving MLIR source locations from IRON/Python through aiecc for better diagnostics → @hunhoffe - Expanded CERT runtime support
…e is now implemented in Rust: Pliron replaces the MLIR-based intermediate representation, and Turso replaces the bundled SQLite dependency. CubeCL also adds LLVM-based CU
…g, and additional operators. NXP Backend finished MLIR based Neutron Converter flow rollout - all operators using the MLIR flow, Neutron backend now support 48 ATen opera
…g, and additional operators. NXP Backend finished MLIR based Neutron Converter flow rollout - all operators using the MLIR flow, Neutron backend now support 48 ATen opera
…ion 3.1.1. eIQ Neutron SDK 3.1.x introduced a new MLIR-based conversion flow with broader operator support for Neutron-C architectures. - Operators migrated to the MLIR f
… [byteir] support byre serialization on cuda byre mlir [byteir] support decomposition of byteir.softmax, byteir.arg max, byteir.addn [byteir] support merge two modules by