The Windows Display Driver Model (WDDM) and the Terminal Control Center (TCC) are two different approaches to managing graphics rendering and display control on Windows operating systems. While both models have their own strengths and weaknesses, WDDM has become the more popular and widely-used display driver model in recent years. In this article, we'll explore the differences between TCC and WDDM, and discuss which one is better.
nvidia-smi -g [GPU_ID] -fdm 0 (You will need to run this in a Command Prompt as Administrator). Will TCC make my games faster?
Do you need a physical monitor or DirectX? │ ├─ Yes → WDDM (only choice) │ └─ No → Do you need Remote Desktop GPU acceleration? │ ├─ Yes → WDDM (RemoteFX / RDP GPU requires WDDM) │ └─ No → Is this a pure compute server? │ └─ Yes → TCC (unquestionably better) tcc wddm better
TCC (Tesla Compute Cluster) offers superior performance for high-performance computing, deep learning, and multi-GPU scaling by reducing overhead and eliminating display-related constraints, as detailed in NVIDIA's documentation [1]. Conversely, WDDM (Windows Display Driver Model) is the necessary standard for gaming and general Windows desktop use, as it supports display outputs and DirectX, according to Wikipedia [2]. For more details, visit NVIDIA Documentation
Connected to your monitors. Handles Windows, web browsers, video playback, and software viewports. The Windows Display Driver Model (WDDM) and the
There is no “TCC + WDDM” on a single GPU. But on multi-GPU systems, combining + N TCC GPUs for work is the optimal architecture for Windows-based compute servers.
TCC模式默认开启/关闭示例： | 产品系列 | 默认模式 | | :--- | :--- | | (如 K20Xm, M2070) | TCC (默认) | | Quadro 系列 (Kepler/Maxwell/Pascal) | WDDM (默认) | | GeForce 系列 | 仅WDDM | nvidia-smi -g [GPU_ID] -fdm 0 (You will need
Conversely, , run user interfaces, or interact with graphics APIs like DirectX and OpenGL.
WDDM is great for —designers, engineers, gamers, or anyone running a GUI.
Choose if you are running a workstation dedicated to AI/HPC and you are running the card in a "headless" (no monitor) configuration.
For thousands of small kernel launches (common in deep learning or physics simulations), this overhead can reduce effective throughput by .
