Skip to content
CUDA 中文指南
搜索文档
K
Main Navigation
阅读指南
全部章节
学习笔记
外观
章节目录
返回顶部
本页小节
4.18 Compute Fabric Transport
待翻译
本节已建立目录框架,正文尚未翻译。
4.18.1 Prerequisites and Scope
4.18.2 Preliminaries
4.18.2.1 Definitions
4.18.2.2 Query for Support
4.18.3 API Overview
4.18.4 Verify clique membership
4.18.5 Creating an Endpoint
4.18.6 Logical Endpoint Ids
4.18.6.1 Limits and Alignment
4.18.7 Sharing an Endpoint
4.18.7.1 Confirming Readiness
4.18.8 Binding Memory
4.18.9 Fabric Operations
4.18.9.1 Per-Thread and Warp-Collective Operations
4.18.10 Completion Status and Error Reporting
4.18.10.1 Configuring Barrier Arrivals and Transaction Counts
4.18.11 Cleaning Up