AI Cloud Computing Platform Solution

Fast, stable, elastic GPU resources with dataset and model management for AI cloud platforms.

Powerful Computing Capability

通过构建云GPU计算资源,具有高并行、高吞吐、低时延 等特点,性能比传统架构显著提高

Containerized Deep Learning Platform

容器化学习平台实现快速环境部署及作业指派,有效降低 多应用程序安装、迁移复杂度

Serve Both AI and HPC Computing Environments

支持SLURM和K8S双调度系统,及主流计算框架,灵活的 资源调度策略,使训练过程更加高效、弹性

Solution Architecture

人工智能云计算平台解决方案.jpg

Sugon’s AI cloud-computing platform solution relies on elastic cloud resources at the bottom and, in the middle, uses XMachine high-performance clusters to form a cloud GPU computing platform. Compute data, logs, and model data are stored uniformly in Sugon ParaStor300 distributed parallel storage. Through Sugon’s containerized enterprise distributed deep-learning platform, it provides dataset management, model management, training, and other services. It helps users with multi-group resource allocation, rapid development-environment setup, and flexible application migration. Based on mainstream deep-learning frameworks, it supports Caffe, TensorFlow, and more, combined with Kubernetes and Docker, providing experiment environments, offline jobs, and online services, connecting the full process from model research and batch training to online prediction, perfectly meeting one-stop deep-learning training and real-time inference, and is currently a high cost-performance overall AI training and inference solution.

Solution Value

Recommended Products

ParaStor300S
ParaStor300S is Sugon’s scale-out high-end distributed storage for massive unstructured data in the big-data era. It is widely used in bioinformatics, weather forecasting, environmental monitoring, seismic monitoring, energy exploration, and video surveillance.
View Details >

Contact Us

After-Sales Service

Solemn Statement