Cadence Design Systems, Inc. has launched the Tensilica HiFi iQ DSP IP, marking the sixth generation of its HiFi DSP family, designed for advanced voice AI and immersive audio applications. This new architecture greatly enhances performance and energy efficiency for modern SoCs, applicable in home entertainment, automotive infotainment, and smartphones.
Compared to its predecessor, the Tensilica HiFi 5s DSP, the HiFi iQ DSP offers twice the raw compute performance, up to eight times the AI performance, and over 25% energy savings across typical workloads. It supports advanced audio codecs and AI-driven tasks like keyword spotting and active noise cancellation, making it suitable for next-generation applications that require sophisticated processing under strict power constraints.
Expected to reach lead customers by early 2026, the HiFi iQ DSP includes improvements in auto-vectorization and energy efficiency, supporting small and large language models directly on the DSP. The product will also integrate with Cadence Neo NPUs and provides software support via the Cadence NeuroWeave SDK and TensorFlow Lite for Micro, facilitating quick deployment for developers.
