Expedera provides high performance and energy-efficient Neural Processing Unit (NPU) IP.
The deep learning accelerator IP uses a unified compute pipeline architecture that enables highly efficient hardware scheduling and advanced memory management. The architecture also reduces memory requirements, allowing high-resolution NN processing, such as 4K/8K video, to run real-time on-chip.
Its product range covers image-related tasks, natural language processing, and audio noise reduction in applications from power-sensitive edge nodes and smartphones to performance-focused automotive and data centers.
