GPU Tile 1 Tile 0 Xe LinkProject Goals - Offer high-level, standard C++ distributed data structures - Support distributed algorithms - Achieve high performance for both multi-GPU, NUMA, and multi-node reduce(par_unseq, z, 0, std::plus()); }Outline - Background (Ranges, Parallelism, Distributed Data Structures) - Distributed Ranges (Concepts) - Implementation (Algorithms and views) - Complex sparse matrices) - Lessons learnedOutline - Background (Ranges, Parallelism, Distributed Data Structures) - Distributed Ranges (Concepts) - Implementation (Algorithms and views) - Complex

program for a supercomputer? Introduce PGAS Model, RDMA Building Remote Pointer Types Building Distributed Data Structures Extending to GPUsThis Talk Background: how do we write a program for a supercomputer supercomputer? Introduce PGAS Model, RDMA Building Remote Pointer Types Building Distributed Data Structures Extending to GPUsThis Talk Background: how do we write a program for a supercomputer? Introduce Introduce PGAS Model, RDMA Building Remote Pointer Types Building Distributed Data Structures Extending to GPUsThis Talk Background: how do we write a program for a supercomputer? Introduce PGAS Model

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 4.19. Distributed Transactions with JTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . fee for such copies and further provided that each copy contains this Copyright Notice, whether distributed in print or electronically. 2 Chapter 1. Spring Boot Documentation This section provides a brief fee for such copies and further provided that each copy contains this Copyright Notice, whether distributed in print or electronically. 1.2. Getting Help If you have trouble with Spring Boot, we would

正在进行中 7 4 快速安装 9 5 编译和安装 11 6 客户端工具 23 7 案例应用-线性回归算法测试 29 8 案例应用-逻辑回归算法测试 35 9 部署架构 39 10 Distributed AI 41 11 XuperDB 49 12 Crypto 53 13 我们的团队 57 14 参与开发 & 测试 59 15 参考文献 61 i ii CHAPTER1 �→根 据 配 置 文 件 的 不 同 有 数 据 持 有 节 点 或 存 储 节 点 两 种 类 型 └── xdb-cli //数 据 存 储 服 务 操 作 客 户 端 3. 编译 Distributed AI $ cd ../dai $ make 编译产出为 output 文件夹，内容为： ├── conf │ ├── config-cli.toml //客 户 端 配 置,也 可 xdb-cli 客户端执行如下命令，请求数据持有节点查看存储节点的在线状态： $ ./xdb-cli nodes list --host http://127.0.0.1:8121 3. 部署 Distributed AI 一般多方安全计算至少由两个任务执行节点，所以这里部署两个任务执行节点。每一个任务执行节点 都有一对公私钥，用来标识节点的账户，公私钥可以通过如下 executor-cli 的客户端命令进行获取：

org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express

正在进行中 5 4 快速安装 7 5 编译和安装 9 6 客户端工具 21 7 案例应用-线性回归算法测试 27 8 案例应用-逻辑回归算法测试 33 9 部署架构 37 10 Distributed AI 39 11 XuperDB 41 12 Crypto 45 13 我们的团队 49 14 参与开发 & 测试 51 15 参考文献 53 i ii CHAPTER1 �→根 据 配 置 文 件 的 不 同 有 数 据 持 有 节 点 或 存 储 节 点 两 种 类 型 └── xdb-cli //数 据 存 储 服 务 操 作 客 户 端 3. 编译 Distributed AI $ cd ../dai $ make 编译产出为 output 文件夹，内容为： ├── conf │ ├── config-cli.toml //客 户 端 配 置,也 可 xdb-cli 客户端执行如下命令，请求数据持有节点查看存储节点的在线状态： $ ./xdb-cli nodes list --host http://127.0.0.1:8121 3. 部署 Distributed AI 一般多方安全计算至少由两个任务执行节点，所以这里部署两个任务执行节点。每一个任务执行节点 都有一对公私钥，用来标识节点的账户，公私钥可以通过如下 executor-cli 的客户端命令进行获取：

org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express

org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express

documentation! 整体介绍 系统介绍 基本概念 正在进行中 快速入门 快速安装 编译和安装 源码编译和安装 通过 docker 安装 客户端工具 操作XuperDB 操作Distributed AI 案例应用-线性回归算法测试 案例简介 测试脚本说明 上传样本文件 训练任务 预测任务 模型评估 案例应用-逻辑回归算法测试 案例简介 测试脚本说明 上传样本文件 训练任务 预测任务 计算需求方(Requester) 任务执行节点(Executor Node) 数据持有节点(DataOwner Node) 存储节点(Storage Node) 区块链节点(Blockchain Node) Distributed AI 服务组件 多方安全计算框架 可信联邦学习 接口与消息定义 配置说明 命令行工具 XuperDB 背景和目标 特点和优势 架构设计 功能介绍 如何使用 Crypto 数据隐私保护 //数据存储服务启动二进制,根据配置文件的不同有数 据持有节点或存储节点两种类型 └── xdb-cli //数据存储服务操作客户端 3. 编译Distributed AI $ cd ../dai $ make 编译产出为 output 文件夹，内容为： ├── conf │ ├── config-cli.toml //客户端配置,也可以使用config

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 25 Multi-processing and Distributed Computing 263 25.1 Code Availability and Loading Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 888 66 Delimited Files 908 67 Distributed Computing 913 67.1 Cluster Manager Interface . . . . . . . . . . . . . . . . . . . . . . . . need to vectorize code for performance; devectorized code is fast • Designed for parallelism and distributed computation • Lightweight "green" threading (coroutines) • Unobtrusive yet powerful type system