Halide学习笔记----Halide tutorial源码阅读11

Halide入门教程11

// Halide tutorial lesson 11: Cross-compilation
// Halide教程第十一课：跨平台编译
// This lesson demonstrates how to use Halide as a cross-compiler to
// generate code for any platform from any platform.
// 本科展示了如何用Halide来充当跨平台编译器来生成不同平台的代码

// On linux, you can compile and run it like so:
// g++ lesson_11*.cpp -g -std=c++11 -I ../include -L ../bin -lHalide -lpthread -ldl -o lesson_11
// LD_LIBRARY_PATH=../bin ./lesson_11

#include "Halide.h"
#include <stdio.h>
using namespace Halide;

int main(int argc, char **argv) {

    // We'll define the simple one-stage pipeline that we used in lesson 10.
    Func brighter;
    Var x, y;

    // Declare the arguments.
    Param<uint8_t> offset;
    ImageParam input(type_of<uint8_t>(), 2);
    std::vector<Argument> args(2);
    args[0] = input;
    args[1] = offset;

    // Define the Func.
    brighter(x, y) = input(x, y) + offset;

    // Schedule it.
    brighter.vectorize(x, 16).parallel(y);

    // The following line is what we did in lesson 10. It compiles an
    // object file suitable for the system that you're running this
    // program on.  For example, if you compile and run this file on
    // 64-bit linux on an x86 cpu with sse4.1, then the generated code
    // will be suitable for 64-bit linux on x86 with sse4.1.
    // 下面这一行做的事情和第十课类似。它编译出一个适应与你当前工作平的object文件。
    // 例如你当前工作在代sse4.1指令的x86 cpu的64位linux系统，它生成适应你平台的object文件
    brighter.compile_to_file("lesson_11_host", args, "brighter");

    // We can also compile object files suitable for other cpus and
    // operating systems. You do this with an optional third argument
    // to compile_to_file which specifies the target to compile for.
    // 也可以编译处适合其他cpu和操作系统平台的object文件。在compile_to_file的target变量中制定
    // 特定的编译目标平台即可

    // Let's use this to compile a 32-bit arm android version of this code:
    // 如下是一个制定安卓os和arm cpu的例子
    Target target;
    target.os = Target::Android; // The operating system
    target.arch = Target::ARM;   // The CPU architecture
    target.bits = 32;            // The bit-width of the architecture
    std::vector<Target::Feature> arm_features; // A list of features to set
    target.set_features(arm_features);
    // We then pass the target as the last argument to compile_to_file.
    // 只需在compile_to_file的最后一个target变量中制定编译的目标平台即可
    brighter.compile_to_file("lesson_11_arm_32_android", args, "brighter", target);

    // And now a Windows object file for 64-bit x86 with AVX and SSE 4.1:
    // windows也类似
    target.os = Target::Windows;
    target.arch = Target::X86;
    target.bits = 64;
    std::vector<Target::Feature> x86_features;
    x86_features.push_back(Target::AVX);
    x86_features.push_back(Target::SSE41);
    target.set_features(x86_features);
    brighter.compile_to_file("lesson_11_x86_64_windows", args, "brighter", target);

    // And finally an iOS mach-o object file for one of Apple's 32-bit
    // ARM processors - the A6. It's used in the iPhone 5. The A6 uses
    // a slightly modified ARM architecture called ARMv7s. We specify
    // this using the target features field.  Support for Apple's
    // 64-bit ARM processors is very new in llvm, and still somewhat
    // flaky.
    // arm/ios也类似
    target.os = Target::IOS;
    target.arch = Target::ARM;
    target.bits = 32;
    std::vector<Target::Feature> armv7s_features;
    armv7s_features.push_back(Target::ARMv7s);
    target.set_features(armv7s_features);
    brighter.compile_to_file("lesson_11_arm_32_ios", args, "brighter", target);


    // Now let's check these files are what they claim, by examining
    // their first few bytes.
    // 下面是通过代码验证编译粗来的object文件是否符合对应文件格式的要求

    // 32-arm android object files start with the magic bytes:
    uint8_t arm_32_android_magic[] = {0x7f, 'E', 'L', 'F', // ELF format
                                      1,       // 32-bit
                                      1,       // 2's complement little-endian
                                      1};      // Current version of elf

    FILE *f = fopen("lesson_11_arm_32_android.o", "rb");
    uint8_t header[32];
    if (!f || fread(header, 32, 1, f) != 1) {
        printf("Object file not generated\n");
        return -1;
    }
    fclose(f);

    if (memcmp(header, arm_32_android_magic, sizeof(arm_32_android_magic))) {
        printf("Unexpected header bytes in 32-bit arm object file.\n");
        return -1;
    }

    // 64-bit windows object files start with the magic 16-bit value 0x8664
    // (presumably referring to x86-64)
    uint8_t win_64_magic[] = {0x64, 0x86};

    f = fopen("lesson_11_x86_64_windows.obj", "rb");
    if (!f || fread(header, 32, 1, f) != 1) {
        printf("Object file not generated\n");
        return -1;
    }
    fclose(f);

    if (memcmp(header, win_64_magic, sizeof(win_64_magic))) {
        printf("Unexpected header bytes in 64-bit windows object file.\n");
        return -1;
    }

    // 32-bit arm iOS mach-o files start with the following magic bytes:
    uint32_t arm_32_ios_magic[] = {0xfeedface, // Mach-o magic bytes
                                   12,  // CPU type is ARM
                                   11,  // CPU subtype is ARMv7s
                                   1};  // It's a relocatable object file.
    f = fopen("lesson_11_arm_32_ios.o", "rb");
    if (!f || fread(header, 32, 1, f) != 1) {
        printf("Object file not generated\n");
        return -1;
    }
    fclose(f);

    if (memcmp(header, arm_32_ios_magic, sizeof(arm_32_ios_magic))) {
        printf("Unexpected header bytes in 32-bit arm ios object file.\n");
        return -1;
    }

    // It looks like the object files we produced are plausible for
    // those targets. We'll count that as a success for the purposes
    // of this tutorial. For a real application you'd then need to
    // figure out how to integrate Halide into your cross-compilation
    // toolchain. There are several small examples of this in the
    // Halide repository under the apps folder. See HelloAndroid and
    // HelloiOS here:
    // https://github.com/halide/Halide/tree/master/apps/
    // 经过上面的验证，halide针对那些目标平台生成的目标文件貌似可信。但是对于一个实际的引用，还必须指出如何
    // 将Halide集成到跨平台的工具链道中。在Halide的github的app中有对应的例子，具体情况参见链接
    printf("Success!\n");
    return 0;
}

在终端中编译和运行代码：

$ g++ lesson_11*.cpp -g -std=c++11 -I ../include -L ../bin -lHalide -lpthread -ldl -o lesson_11
$ ./lesson_11

结果

补充说明：（更具体的情况点链接，去官网文档上看吧，比这里详细）

1.compile_to_file函数

EXPORT void Halide::Func::compile_to_file(
const std::string & filename_prefix,                    // 文件名前缀
const std::vector< Argument > & args,                   // 参数列表
const std::string & fn_name = "",                       // 函数名
const Target &  target = get_target_from_environment()  // 目标平台
)   

2.Target结构

OS	Arch
OSUnknown	ArchUnknown
Linux	X86
Windows	ARM
OSX	MIPS
Android	Hexagon
IOS	POWERPC
QuRT
NoOS

内容太多，不在此一一列举，需要什么点链接上去查即可

http://halide-lang.org/docs/struct_halide_1_1_target.html#details