第 4 章 CUDA 功能
本章正文待翻译,点击下方小节进入对应页面。
本章目录
- 4.1 Unified Memory
- 4.2 CUDA Graphs
- 4.3 Stream-Ordered Memory Allocator
- 4.4 Cooperative Groups
- 4.5 Programmatic Dependent Launch and Synchronization
- 4.6 Green Contexts
- 4.7 Locality Domains
- 4.8 Lazy Loading
- 4.9 Error Log Management
- 4.10 Asynchronous Barriers
- 4.11 Pipelines
- 4.12 Asynchronous Data Copies
- 4.13 Work Stealing with Cluster Launch Control
- 4.14 L2 Cache Control
- 4.15 Memory Synchronization Domains
- 4.16 Interprocess Communication
- 4.17 Virtual Memory Management
- 4.18 Compute Fabric Transport
- 4.19 Extended GPU Memory
- 4.20 CUDA Dynamic Parallelism
- 4.21 CUDA Interoperability with APIs
- 4.22 Driver Entry Point Access