debug的时候需要等很长时间读模型,查资料发现了两种快速读取大文件的方法。
test 1:每次读一个字符串
test 2、3一次读取整个文件
{//test 1
string buf;
clock_t start = clock();
ifstream fin(objpath); while (fin >> buf); fin.close();
clock_t end = clock();
cout << "time : " << ((double)end - start) / CLOCKS_PER_SEC << "s\n";
}
{//test 2
clock_t start = clock();
ifstream fin(objpath, std::ios::binary); vector<char> buf(fin.seekg(, std::ios::end).tellg());
fin.seekg(, std::ios::beg).read(&buf[], static_cast<std::streamsize>(buf.size())); fin.close();
clock_t end = clock();
cout << "time : " << ((double)end - start) / CLOCKS_PER_SEC << "s\n";
}
{//test 3
clock_t start = clock();
ifstream fin(objpath); stringstream buf;
buf << fin.rdbuf(); fin.close();
clock_t end = clock();
cout << "time : " << ((double)end - start) / CLOCKS_PER_SEC << "s\n";
}
文件大小为112M,花费的时间分别为:
time : .809s
time : .092s
time : .213s
于是将loader改成了第二种。