Is this your business? Claim it to manage your IP and profile
Targeted at extremely demanding AI applications, the Origin E8 NPU IP core provides unrivaled processing capabilities, scaling up to 128 TOPS in a single core. It is well-suited for automotive ADAS systems and data center operations, capable of running multiple networks simultaneously with efficient resource use. The system's architecture allows unparalleled performance without the latency issues associated with traditional solutions, ensuring consistently high output with low power requirements.
The Origin E2 NPU core is designed for efficiency in power and area, ideal for smartphones and similar edge devices. It offers flexible configurations from 1 to 20 TOPS, supporting a range of neural networks including CNNs and RNNs. By maintaining high processor utilization and minimizing latency, the E2 delivers optimal performance for both consumer and industrial applications. Its native support for multiple network architectures ensures that it can handle demanding AI workloads with minimal resource consumption.
The Origin E1 neural engine is tailored to deliver optimized performance for networks in devices that demand low power and minimal external memory usage, such as home appliances and smartphones. It features the LittleNPU architecture, which caters to always-on applications by efficiently processing data with minimal power, often running as low as 10-20 mW. The unique packet-based architecture allows for parallel execution across multiple neural network layers, enhancing performance while keeping power usage and latency low.
The TimbreAI T3 is optimized for audio applications in power-constrained environments, offering an ultra-low-power AI interface for noise reduction in devices like headsets. With a performance of 3.2 GOPS at less than 300 µW, it is designed to operate efficiently without requiring external memory. It supports common audio neural network models while maintaining the accuracy and performance of trained models without additional optimizations.
Designed for high-efficiency AI processing, the Origin E6 supports up to 32 TOPS, making it suitable for next-generation devices like AR/VR headsets. It excels at balancing performance with power usage, supporting both traditional and generative AI models across various device applications. The packet-based architecture allows for seamless integration of complex AI models, ensuring high processor utilization without the need for hardware-specific optimizations.
Join the world's most advanced semiconductor IP marketplace!
It's free, and you'll get all the tools you need to evaluate IP, download trial versions and datasheets, and manage your evaluation workflow!
To evaluate IP you need to be logged into a buyer profile. Select a profile below, or create a new buyer profile for your company.
