Course Overview
The course provides an overview of the Intel Gaudi AI accelerator, a purpose-built deep learning acceleration processor for both deep learning training and inference at scale. The course covers the hardware and software architecture of the Intel Gaudi AI accelerator, including its matrix multiplication engine, Tensor processing cores, and 96 gigabytes of onboard HBM2E memory. It also discusses the SynapseAI software stack, which is designed for performance and ease of use, and supports PyTorch and TensorFlow models. The course explains how to migrate models from GPUs to the Intel Gaudi AI accelerator using the GPU Migration Toolkit and how to run optimized generative AI and large language models on the Intel Gaudi 2. Additionally, it covers rack-level integration for the Intel Gaudi 2 AI accelerator, including the reference server, connectivity, and administrative tools needed to manage the platform.
Share With
Chapter List
Quiz Status
Please watch the video before taking the quiz.
Once you have watched the video, please take the quiz.
Chapter Overview
The course provides an overview of the Intel Gaudi AI accelerator, a purpose-built deep learning acceleration processor for both deep learning training and inference at scale. The course covers the hardware and software architecture of the Intel Gaudi AI accelerator, including its matrix multiplication engine, Tensor processing cores, and 96 gigabytes of onboard HBM2E memory. It also discusses the SynapseAI software stack, which is designed for performance and ease of use, and supports PyTorch and TensorFlow models. The course explains how to migrate models from GPUs to the Intel Gaudi AI accelerator using the GPU Migration Toolkit and how to run optimized generative AI and large language models on the Intel Gaudi 2. Additionally, it covers rack-level integration for the Intel Gaudi 2 AI accelerator, including the reference server, connectivity, and administrative tools needed to manage the platform.