2.5 检测处理器指令集
NOTE:此示例代码可以在 https://github.com/dev-cafe/cmake-cookbook/tree/v1.0/chapter-02/recipe-05 中找到,包含一个C++示例。该示例在CMake 3.10版(或更高版本)中是有效的,并且已经在GNU/Linux、macOS和Windows上进行过测试。
本示例中,我们将讨论如何在CMake的帮助下检测主机处理器支持的指令集。这个功能是较新版本添加到CMake中的,需要CMake 3.10或更高版本。检测到的主机系统信息,可用于设置相应的编译器标志,或实现可选的源代码编译,或根据主机系统生成源代码。本示例中,我们的目标是检测主机系统信息,使用预处理器定义将其传递给C++源代码,并将信息打印到输出中。
准备工作
我们是C++源码(processor-info.cpp)如下所示:
#include "config.h"#include <cstdlib>#include <iostream>int main(){std::cout << "Number of logical cores: "<< NUMBER_OF_LOGICAL_CORES << std::endl;std::cout << "Number of physical cores: "<< NUMBER_OF_PHYSICAL_CORES << std::endl;std::cout << "Total virtual memory in megabytes: "<< TOTAL_VIRTUAL_MEMORY << std::endl;std::cout << "Available virtual memory in megabytes: "<< AVAILABLE_VIRTUAL_MEMORY << std::endl;std::cout << "Total physical memory in megabytes: "<< TOTAL_PHYSICAL_MEMORY << std::endl;std::cout << "Available physical memory in megabytes: "<< AVAILABLE_PHYSICAL_MEMORY << std::endl;std::cout << "Processor is 64Bit: "<< IS_64BIT << std::endl;std::cout << "Processor has floating point unit: "<< HAS_FPU << std::endl;std::cout << "Processor supports MMX instructions: "<< HAS_MMX << std::endl;std::cout << "Processor supports Ext. MMX instructions: "<< HAS_MMX_PLUS << std::endl;std::cout << "Processor supports SSE instructions: "<< HAS_SSE << std::endl;std::cout << "Processor supports SSE2 instructions: "<< HAS_SSE2 << std::endl;std::cout << "Processor supports SSE FP instructions: "<< HAS_SSE_FP << std::endl;std::cout << "Processor supports SSE MMX instructions: "<< HAS_SSE_MMX << std::endl;std::cout << "Processor supports 3DNow instructions: "<< HAS_AMD_3DNOW << std::endl;std::cout << "Processor supports 3DNow+ instructions: "<< HAS_AMD_3DNOW_PLUS << std::endl;std::cout << "IA64 processor emulating x86 : "<< HAS_IA64 << std::endl;std::cout << "OS name: "<< OS_NAME << std::endl;std::cout << "OS sub-type: "<< OS_RELEASE << std::endl;std::cout << "OS build ID: "<< OS_VERSION << std::endl;std::cout << "OS platform: "<< OS_PLATFORM << std::endl;return EXIT_SUCCESS;}
其包含config.h头文件,我们将使用config.h.in生成这个文件。config.h.in如下:
#pragma once#define NUMBER_OF_LOGICAL_CORES @_NUMBER_OF_LOGICAL_CORES@#define NUMBER_OF_PHYSICAL_CORES @_NUMBER_OF_PHYSICAL_CORES@#define TOTAL_VIRTUAL_MEMORY @_TOTAL_VIRTUAL_MEMORY@#define AVAILABLE_VIRTUAL_MEMORY @_AVAILABLE_VIRTUAL_MEMORY@#define TOTAL_PHYSICAL_MEMORY @_TOTAL_PHYSICAL_MEMORY@#define AVAILABLE_PHYSICAL_MEMORY @_AVAILABLE_PHYSICAL_MEMORY@#define IS_64BIT @_IS_64BIT@#define HAS_FPU @_HAS_FPU@#define HAS_MMX @_HAS_MMX@#define HAS_MMX_PLUS @_HAS_MMX_PLUS@#define HAS_SSE @_HAS_SSE@#define HAS_SSE2 @_HAS_SSE2@#define HAS_SSE_FP @_HAS_SSE_FP@#define HAS_SSE_MMX @_HAS_SSE_MMX@#define HAS_AMD_3DNOW @_HAS_AMD_3DNOW@#define HAS_AMD_3DNOW_PLUS @_HAS_AMD_3DNOW_PLUS@#define HAS_IA64 @_HAS_IA64@#define OS_NAME "@_OS_NAME@"#define OS_RELEASE "@_OS_RELEASE@"#define OS_VERSION "@_OS_VERSION@"#define OS_PLATFORM "@_OS_PLATFORM@"
如何实施
我们将使用CMake为平台填充config.h中的定义,并将示例源文件编译为可执行文件:
首先,我们定义了CMake最低版本、项目名称和项目语言:
cmake_minimum_required(VERSION 3.10 FATAL_ERROR)project(recipe-05 CXX)
然后,定义目标可执行文件及其源文件,并包括目录:
add_executable(processor-info "")target_sources(processor-infoPRIVATEprocessor-info.cpp)target_include_directories(processor-infoPRIVATE${PROJECT_BINARY_DIR})
继续查询主机系统的信息,获取一些关键字:
foreach(keyIN ITEMSNUMBER_OF_LOGICAL_CORESNUMBER_OF_PHYSICAL_CORESTOTAL_VIRTUAL_MEMORYAVAILABLE_VIRTUAL_MEMORYTOTAL_PHYSICAL_MEMORYAVAILABLE_PHYSICAL_MEMORYIS_64BITHAS_FPUHAS_MMXHAS_MMX_PLUSHAS_SSEHAS_SSE2HAS_SSE_FPHAS_SSE_MMXHAS_AMD_3DNOWHAS_AMD_3DNOW_PLUSHAS_IA64OS_NAMEOS_RELEASEOS_VERSIONOS_PLATFORM)cmake_host_system_information(RESULT _${key} QUERY ${key})endforeach()
定义了相应的变量后,配置
config.h:configure_file(config.h.in config.h @ONLY)
现在准备好配置、构建和测试项目:
$ mkdir -p build$ cd build$ cmake ..$ cmake --build .$ ./processor-infoNumber of logical cores: 4Number of physical cores: 2Total virtual memory in megabytes: 15258Available virtual memory in megabytes: 14678Total physical memory in megabytes: 7858Available physical memory in megabytes: 4072Processor is 64Bit: 1Processor has floating point unit: 1Processor supports MMX instructions: 1Processor supports Ext. MMX instructions: 0Processor supports SSE instructions: 1Processor supports SSE2 instructions: 1Processor supports SSE FP instructions: 0Processor supports SSE MMX instructions: 0Processor supports 3DNow instructions: 0Processor supports 3DNow+ instructions: 0IA64 processor emulating x86 : 0OS name: LinuxOS sub-type: 4.16.7-1-ARCHOS build ID: #1 SMP PREEMPT Wed May 2 21:12:36 UTC 2018OS platform: x86_64
输出会随着处理器的不同而变化。
工作原理
CMakeLists.txt中的foreach循环会查询多个键值,并定义相应的变量。此示例的核心函数是cmake_host_system_information,它查询运行CMake的主机系统的系统信息。本例中,我们对每个键使用了一个函数调用。然后,使用这些变量来配置config.h.in中的占位符,输入并生成config.h。此配置使用configure_file命令完成。最后,config.h包含在processor-info.cpp中。编译后,它将把值打印到屏幕上。我们将在第5章(配置时和构建时操作)和第6章(生成源代码)中重新讨论这种方法。
更多信息
对于更细粒度的处理器指令集检测,请考虑以下模块: https://github.com/VcDevel/Vc/blob/master/cmake/OptimizeForArchitecture.cmake 。有时候,构建代码的主机可能与运行代码的主机不一样。在计算集群中,登录节点的体系结构可能与计算节点上的体系结构不同。解决此问题的一种方法是,将配置和编译作为计算步骤,提交并部署到相应计算节点上。
