Docs

GPU Compute Service

Overview

Applicable scenarios

GPU instances are suitable for scenarios requiring large-scale parallel computing, such as AI deep learning, image/video rendering, and scientific computing.

AI deep learning

Model training scenarios

GPU graphics cards such as the V-Series and A-Series GPU graphics cards are suitable for AI model training. They offer large memory capacity and high-speed access capabilities, combined with NVLink multi-GPU interconnect technology, providing exceptional computing power for multi-GPU parallel processing.

Inference scenarios

GPU graphics cards such as the T-Series and A-Series GPU graphics cards deliver acceleration with high performance-to-power ratio for AI inference, and are widely used in image recognition, language translation, and similar scenarios.

Image/video rendering

Heterogeneous GPU rendering instances provide high-performing 3D graphics virtualization for efficient real-time image rendering. This type is widely used in fields such as engineering graphics, video rendering, and animation production, significantly increasing real-time and offline rendering efficiency.

Last updated: 2025.11.21 16:58:17