OpenCL* on Intel® Programmable Acceleration Card with Intel® Arria® 10 GX FPGA快速入门用户指南

下载 PDF
ID 683831
日期 12/04/2018
Public
文档目录
文档目录 x
1. 关于本文档 2. 引言 3. 设置主机(Setting Up the Host Machine) 4. 运行诊断程序(Running Diagnostics) 5. 对多卡系统的 OpenCL* 支持 6. 运行示例(Running Samples) 7. 编译OpenCL内核(Compiling OpenCL Kernels) 8. 运行一个OpenCL设计实例 9. OpenCL* on the Intel® PAC with Intel® Arria® 10 GX FPGA快速入门用户指南存档 10. OpenCL* on Intel® Programmable Acceleration Card with Intel® Arria® 10 GX FPGA快速入门用户指南的文档修订历史 11. 禁用非统一存储器访问(NUMA)和DMA工作线程以优化 PCIe* 带宽
1. 关于本文档 x
1.1. 约定 1.2. 加速术语表(Acceleration Glossary) 1.3. 略缩语
2. 引言 x
2.1. 发布内容
3. 设置主机(Setting Up the Host Machine) x
3.1. 安装发行版(Installing the Release) 3.2. 初始化Intel Acceleration Stack for OpenCL* 3.3. 配置OpenCL Driver
6. 运行示例(Running Samples) x
6.1. 运行Hello World 6.2. 运行矢量加(Vector Add)
7. 编译OpenCL内核(Compiling OpenCL Kernels) x
7.1. 检查时序结果
8. 运行一个OpenCL设计实例 x
8.1. 下载 OpenCL* 设计实例 8.2. 编译内核(Compiling the Kernel) 8.3. 编译主机(Compiling the Host) 8.4. 运行可执行文件(Running the Executable)
1. 关于本文档
2. 引言
3. 设置主机(Setting Up the Host Machine)
4. 运行诊断程序(Running Diagnostics)
5. 对多卡系统的 OpenCL* 支持
6. 运行示例(Running Samples)
7. 编译OpenCL内核(Compiling OpenCL Kernels)
8. 运行一个OpenCL设计实例
9. OpenCL* on the Intel® PAC with Intel® Arria® 10 GX FPGA快速入门用户指南存档
10. OpenCL* on Intel® Programmable Acceleration Card with Intel® Arria® 10 GX FPGA快速入门用户指南的文档修订历史
11. 禁用非统一存储器访问(NUMA)和DMA工作线程以优化 PCIe* 带宽

8.4. 运行可执行文件(Running the Executable)

在运行主机(host)前，必须要有已编译的 OpenCL* kernel和host。
要在硬件上运行主机程序，需输入以下命令来运行bin目录下名为host的可执行文件： 
$ ./bin/host
输出显示类似如下： 
Using AOCX:fft2d.aocx

Reprogramming device [0] with handle 1
Launching FFT transform (ordered data layout)
Kernel initialization is complete.
	Processing time= 3.0598ms
	Throughput= 0.3427 Gpoints/sec (34.2690 Gflops)
	Signal to noise ratio on output sample: 137.231003 → PASSED

Launching inverse FFT transform (ordered data layout)
Kernel initialization is complete.
	Processing time= 3.0628ms
	Throughput= 0.3424 Gpoints/sec (34.2364 Gflpos)
	Signal to noise ration on output sample: 136.860797 → PASSED

Launching FFT transform (alternative data layout)
Kernel initialization is complete.
	Processing time= 2.2904ms
	Throughput= 0.4578 Gpoints/sec (45.7821 Gflops)
	Signal to noise ration on output sample: 137.435876 → PASSED

Launching inverse FFT transform (alternative data layout)
Kernel initialization is complete.
	Processing time= 2.3253ms
	Throughput= 0.4509 Gpoints/sec (45.0934 Gflops)
	Signal to noise ration on output sample: 136.689050 → PASSED
二级标题