Current location - Quotes Website - Team slogan - What is the difference between TCC mode and WDDM mode of Nvidia GPU?
What is the difference between TCC mode and WDDM mode of Nvidia GPU?
NVIDIA high-performance GPUs (such as Tesla and Quadro) can be configured in TCC(Tesla Computing Cluster) mode or WDDM(Windows Display Driven Model) mode.

The difference between the two is that in TCC mode, GPU is completely used for calculation and does not need local display function; In WDDM mode, GPU also plays the role of computing card+graphics card.

Use:

Operation of cuda program on remote desktop under 1.Windows system

2. The running of cuda cluster program in GPU cluster environment under 2.Windows system.

The full name of TCC Tesla Compute Cluster is one of two driving modes of Nvidia graphics card, and the other mode is WDDM (default mode, full name of Windows display driving model).

Use nvidia-smi tool to open TCC mode. Nvidia-smi is a system management interface provided by NVIDIA, which is located in the NVSMI folder under the root directory of driver installation, and its full name is nvidia-smi.exe. Of course, the main function of this tool is not only to open TCC, but also to manage graphics cards. For other functions of this tool, you can directly enter -h to view. You usually need a c++ compiler to add cl.exe to the path.

Graphics cards that support this tool:

Tesla: S 1070, S2050, C 1060, C2050/70/75, M2050/70/90, X2070/90.

Quadro: 4000, 5000, 6000, 7000 and M2070-Q.

Open command: nvidia-smi -dm 1+0. If there is no such internal command, add the path of nvidia-smi.exe to the environment variable.