http://gmd20.blog.163.com/blog/static/168439232012113111759514/
執行 10000000 次, 耗時 2258,369 微秒 QueryPerformanceCounter
執行 10000000 次, 耗時 26,347 微秒 GetTickCount
執行 10000000 次, 耗時 242,879 微秒 time()
c的時間函式 time(time_t) 大概比GetSystemTimeAsFileTime慢6倍,比_ftime 快6倍
執行 10000000 次, 耗時 1310,066 微秒 _ftime
執行 10000000 次, 耗時 1722,125 微秒 GetLocalTime
執行 10000000 次, 耗時 39,131 微秒 GetSystemTimeAsFileTime
GetLocalTime耗時等於 = GetSystemTimeAsFileTime 耗時+ FileTimeToSystemTime 的耗時
------------
可以看到精度越高效能越差
GetTickCount 精度1毫秒 > GetLocalTime 精度100納秒 (0.1 微秒) > QueryPerformanceCounter (搞不懂這個怎麼這麼差)
如果僅僅為了計算時間偏差,可以使用 GetSystemTimeAsFileTime,這個精度可以達到100納秒,
msdn有個介紹。
http://msdn.microsoft.com/ZH-CN/library/windows/desktop/ms724284(v=vs.85).aspx
Contains a 64-bit value representing the number of 100-nanosecond intervals since January 1, 1601 (UTC).
It is not recommended that you add and subtract values from the FILETIME structure to obtain relative times. Instead, you should copy the low- and high-order parts of the file time to a ULARGE_INTEGER structure, perform 64-bit arithmetic on the QuadPart member, and copy the LowPart and HighPart members into the FILETIME structure.
Do not cast a pointer to a FILETIME structure to either a ULARGE_INTEGER* or __int64* value because it can cause alignment faults on 64-bit Windows.
測試程式碼如下
#include <iomanip>
#include <fstream>
#include <iostream>
#include <map>
#include <sstream>
#include <list>
#include <vector>
#include <stdlib.h>
#include <stdint.h>
#include <stdio.h>
#include <sys/types.h>
#include <sys/timeb.h>
#include <time.h>
#include <Windows.h>
#include "Trace.h"
using
namespace
std;
int
main (
int
,
char
**)
{
LARGE_INTEGER freq, t0, t1;
QueryPerformanceFrequency(&freq);
size_t
number = 10000000;
int
total_counter = 0;
//LARGE_INTEGER t3;
//struct timeb timebuffer;
SYSTEMTIME lt;
FILETIME SystemTimeAsFileTime;
QueryPerformanceCounter(&t0);
for
(
int
i=0; i< number; i++) {
//QueryPerformanceCounter(&t3);
//total_counter += t3.LowPart;
//total_counter += GetTickCount();
//ftime(&timebuffer);
//total_counter += timebuffer.time;
//GetLocalTime(<);
//total_counter += lt.wMilliseconds;
// total_counter += _time32(NULL); time(NULL)
GetSystemTimeAsFileTime(&SystemTimeAsFileTime);
FileTimeToSystemTime(&SystemTimeAsFileTime,<);
total_counter += lt.wMilliseconds;
}
QueryPerformanceCounter(&t1);
int
time
= (((t1.QuadPart-t0.QuadPart)*1000000)/freq.QuadPart);
std::cout <<
"執行 "
<< number <<
" 次, 耗時 "
<<
time
<<
" 微秒"
<< std::endl;
std::cout << total_counter;
int
a;
cin >> a;
return
0;
}
c語言精確到微妙 GetSystemTimeAsFileTime
c語言庫函式中的clock()函式只能精確到ms,若想更精確的us,在網路上查了一遍,完整的可行的解決方案繁瑣,實在沒時間去仔細琢磨。不過找到了一個簡潔的方案:呼叫GetSystemTimeAsFileTime函式,單位是100ns。
1秒=1,000,000 微秒(μs) 1微秒=1/1,000,000秒(s)
1秒=1,000,000,000 納秒(ns) 1納秒=1/1,000,000,000秒(s)
1秒=1,000,000,000,000 皮秒(ps) 1皮秒=1/1,000,000,000,000秒(s)