Clock() 在调试版本上运行良好，但在发布版本上不起作用（C++ VS2010）
这是我的代码：
// Start performance test clock
assert((start=clock())!=-1);
// Some reading and writing methods
// Get stop time
stop = clock();
cout << stop << endl;
// Calculate operation time
double result = (double)(stop-start)/CLOCKS_PER_SEC;
// Print result
cout << "--> Finished analysing in " << result << "s" << endl;
当我调试程序时它效果很好，但是当我运行发布版本时， stop 接收的值比 start 小得多，结果是负数。
有什么想法吗？
This is my code:
// Start performance test clock
assert((start=clock())!=-1);
// Some reading and writing methods
// Get stop time
stop = clock();
cout << stop << endl;
// Calculate operation time
double result = (double)(stop-start)/CLOCKS_PER_SEC;
// Print result
cout << "--> Finished analysing in " << result << "s" << endl;
It works great when I debug my program, but when I run the release version,
stop receives a much smaller value than start, and result is a negative number.
Any ideas?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
start的赋值不应出现在assert语句中。assert在发布版本中通常是无操作。因此，在调试版本中，将执行start=clock()语句，但在发布版本中则不会执行。因此它可能没有初始化（取决于之前的代码和start的声明）。通常希望避免使用有副作用的assert语句；它可能会导致调试和发布版本之间出现细微的差异/错误。像这样写可能会更好：
The assignment of
startshould not be in theassertstatement.assertis typically a no-op in release builds. So in the debug build, the statementstart=clock()would be executed, but in the release build, it would not. So it is possible (depending on earlier code and the declaration ofstart) that it is not initialized. It is typically desirable to avoid usingassertstatements that have side effects; it can lead to subtle differences/bugs between debug and release builds.It might be better to write it like this: