AI Acceleration Whitepaper

Artificial Intelligence (AI) is present in every aspect of our life. AI acceleration technologies have also become more widespread in various industries and vertical technical domains. From the perspective of AI applications, AI technologies are divided into the infrastructure layer, algorithm layer, vertical realm technology layer, and industry-specific solution layer.

This whitepaper focuses on the AI infrastructure layer and how it powers AI Acceleration in the following subordinate layers with specific acceleration solutions based on AI application characteristics in various business scenarios: 

• GPU Resource Layer 
• Resource Scheduling Layer 
• AI Data Acceleration Layer
• AI Computing Acceleration Layer
• AI Job Scheduling Layer 

This whitepaper also discusses the cloud-native AI suite, which integrates a full range of optimization capabilities of all layers and provides powerful tools for AI training management and system monitoring.