8月24日,在小米玄戒芯片技术沟通会上,小米正式推出大模型专用AI加速芯片玄戒O100,采用6nm 3D晶圆级堆叠方案,带宽达1.22TB/s。架构方面,玄戒O100采用近存AI计算架构,搭配玄戒高宽带矩阵总线,支持多核并行计算,运行Xiaomi MiMo 3B模型可实现330Tokens/s推理速度。
特别声明:以上内容(如有图片或视频亦包括在内)为自媒体平台“网易号”用户上传并发布,本平台仅提供信息存储服务。
Notice: The content above (including the pictures and videos if any) is uploaded and posted by a user of NetEase Hao, which is a social media platform and only provides information storage services.