Author: Lex Loep
Date: 22:25:11 01/13/00
Go up one level in this thread
On January 13, 2000 at 19:01:08, Tom Kerrigan wrote: >On January 13, 2000 at 18:48:55, Christophe Theron wrote: > >>Check this again Tom. Lex (author of Rebel-Tiger GUI) told me the same some time >>ago. We both checked and realized that GetTickCount() has the same ~50ms >>accuracy under Windows 95/98. > >No, I'm quite sure of this. > >I use GetTickCount() to compute how fast my program solves test suite problems >(among other things). Here are my solution times from the first few WAC >problems: > >0.869 >(didn't solve) >0.021 >0.010 >0.012 >0.019 >0.021 >0.039 >0.013 >0.017 > >So it seems to me that GetTickCount() is usually accurate to the millisecond. > >It would be trivial to write a program to verify this fact, to be 100% sure. > >-Tom // ticktest.cpp : Defines the entry point for the console application. // Try this: #include "stdafx.h" int main(int argc, char* argv[]) { DWORD t1,t2; printf("Tick test!\n"); for(int i=0; i<5; i++) { Sleep(0); SetThreadPriority(GetCurrentThread(),THREAD_PRIORITY_TIME_CRITICAL); t1=GetTickCount(); while((t2=GetTickCount())==t1); SetThreadPriority(GetCurrentThread(),THREAD_PRIORITY_NORMAL); printf("Tdiff=%d\n",t2-t1); } return 0; } Under W95 resulst are random, under NT diff between two measurement is 10 ms For accurate measurement of small times use QueryPerformanceCounter and QueryPerformanceFrequency Lex
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.