NVIDIA Instances

Last updated: 2020-02-26 17:10:49


Instance Overview

NVIDIA series GPU instance GN * suitable not only for GPU general computing scenarios such as deep learning and scientific computing, but also for graphics and image processing (3D rendering, video codec) scenes; Tencent Cloud and CVMConsistent management style To provide fast, stable and flexible computing services


It is suitable for working scenarios where data throughput is large and calculation speed is demanding.
Deep Learning

  • Graphic and image processing
  • Video codec
  • Graphic database
  • High-performance database
  • Computational fluid dynamics
  • Computational finance
  • Seismic analysis
  • Molecular modeling
  • Genomics and others;


  • GPU performance: The main indicator is GPU's floating-point computing performance. TF stands for T Flops, SP for single-precision floating-point computing, DP for double-precision floating-point computing, INT8 for INT8 integer computing, and DL for Deep Learning Tensor Core computing (V100 only).
  • Storage/network: The storage list shows the storage types supported by the current instance; the network bandwidth refers to the network bandwidth of the physical server where an instance of this type is located. See the purchase page for the network bandwidth assigned by an instance of a certain type.
  • VGPU The GN10X,GN7 instance cluster provides instance types that support vGPU. The type of vGPU is vComputeServer, which only supports CUDA computing API, and does not support graphical API such as DirectX and OpenGL.
  • Availability zone Bei 4 represents Beijing 4, Xinyi represents Singapore 1, and so on.

GN2,GN8 instances provide SSD-based local storage (the storage of GN2 instances forces the selection of fixed-capacity SSD local disks by default. For more information, please see the purchase page). When using local storage, the system disk and data disk of these instances only exist during the instance life cycle; when the instance expires or you take the initiative to Terminate instance, the application and data in its instance storage will be erased. We recommend that you back up the data stored in the instance store on a regular basis.

Selection recommendation

Tencent Cloud provides a wide range of GPU computing examples to meet the needs of different business application scenarios.

Service Options