SemiVoice

anysilicon

06/10/2025, 08:06 AM UTC

芯原股份超低功耗NPU为移动端设备大语言模型推理提供超40TOPS算力VeriSilicon’s Ultra-Low Energy NPU Provides Over 40 TOPS for On-Device LLM Inference in Mobile Applications

➀ 芯原股份NPU IP支持大语言模型端侧推理，算力超40TOPS；

➁ 面向移动应用优化能效设计；

➂ 满足生成式AI在终端设备中日益增长的需求

➀ VeriSilicon's NPU IP achieves >40 TOPS for on-device LLM inference;

➁ Targets mobile applications with optimized energy efficiency;

➂ Addresses growing demand for generative AI in edge devices

---

本文由大语言模型（LLM）生成，旨在为读者提供半导体新闻内容的知识扩展（Beta）。