Intel® oneAPI Deep Neural Network Library (oneDNN) is an open-source performance library for deep learning applications. The library includes basic building blocks for neural networks optimized for Intel Architecture Processors and Intel Processor Graphics. oneDNN is intended for deep learning applications and framework developers interested in improving application performance on Intel Architecture Processors and Intel Processor Graphics. Deep learning practitioners should use one of the applications enabled with oneDNN.

oneDNN is distributed as part of Intel® oneAPI DL Framework Developer Toolkit, the Intel oneAPI Base Toolkit, and is available via apt and yum channels.

oneDNN continues to support features currently available with DNNL, including C and C++ interfaces, OpenMP*, Intel oneAPI Threading Building Blocks, and OpenCL™ runtimes. oneDNN introduces SYCL*/DPC++ API and runtime support for the oneAPI programming model.

For the complete list of features, documentation, code samples, and downloads, visit the official Intel oneAPI Deep Neural Network Library website. If you plan to use oneDNN as part of the oneAPI Base Toolkit, consider that premium support is available as a paid option. For the community-supported open-source version, visit the oneDNN GitHub* page.