首页 » 产品 » GPU 算力加速器 » 超微 GPU 服务器 » 超微 NVIDIA H100 80GB HBM3 GPU

超微 NVIDIA H100 80GB HBM3 GPU

Features the NVIDIA Hopper architecture with 80GB HBM3 memory. Engineered for massive AI training and HPC workloads. Includes a Transformer Engine to accelerate LLM training by up to 9X. Applications: LLM training, scientific simulation, and financial risk modeling.

相关产品

快速联系

* 您的数据受到安全保护,仅用于必要的通信。我们不会共享您的详细资料或发送垃圾邮件、, 隐私政策