Intel Sapphire Rapids CPUs & DG2 Discrete GPUs receive full support in OpenVINO 2022.3
Intel introduced OpenVINO, a toolkit (completely open-source) that assists with optimizing and deploying models for deep-learning instances and boosts performance in several machine-learning frameworks, in 2019. OpenVINO has always offered support, but with this new version, the processor series sees complete support for the deep learning toolkit. This further support will allow more inference workloads in the tool to go straight from edge computing to the cloud. OpenVINO also follows the release of Intel’s oneAPI Deep Neural Network Library (oneDNN) 3.0, utilizing several deep learning applications. OpenVINO is part of Intel’s Long-Term Support, which means that the company will continue to develop and grow the toolkit. Currently, Intel promises to continue support over the next two years. On the dGPU front, OpenVINO will be optimizing support for the discrete graphics based on the DG2 GPU architecture featured on the Arc consumer-level products to the company’s Data Center GPU Flex series. The toolkit should also assist with optimizations for Intel’s 12th and 13th Gen Core CPU series. OpenVINO 2022.3 offers:
Broader Model and Hardware Support: Optimize and deploy quickly across an expanded range of deep learning models, including natural language processing or NLP. Access AI acceleration across an expanded range of hardware. Improved API and More Integrations: Adopting and maintaining your code is more straightforward. This version requires fewer code changes, provides more framework integration options, and minimizes conversions. Expanded Model Coverage: See a performance boost quickly with automatic device discovery, load balancing, and dynamic inference parallelism across processors, graphics cards, and more.
One product line from Intel is being dropped from Intel OpenVINO. The Intel Movidius Vision Processing Unit support has been dropped in this and future releases. The Movidius VPU launched in the first quarter of 2020 and assisted with deep learning instances. However, it was not as widely adopted, which is why the product series has been dropped from future releases. The new update for the OpenVINO Runtime is available on the company’s GitHub page. OpenVINO’s documentation provides a complete listing of supported devices. News Sources: Phoronix, OpenVINO GitHub, OpenVINO
