[ofw][patch][WinVerbs tests] Re: fix IPv6related connectionproblem

Diego Guella diego.guella at sircomtech.com
Thu Feb 19 03:06:29 PST 2009


----- Original Message ----- 
>From: Leonid Keller
>Seems like QueryPerformanceCounter is the most trustful thing for now.
>One needs just not to forget that it takes its own time for work. :(

If so, I think the best options are 2 and 3.
Option 2 is actually what MS does this article: http://support.microsoft.com/kb/172338/en-us

t1 = QueryPerformanceCounter();
t2 = QueryPerformanceCounter();
overhead = t2-t1;

I think it makes sense to measure overhead over a long time (loop some 100000+ times measuring overhead), and take the average or 
other statistical function of it.



>Let's formulate our problem:
>We have to measure a very fast process and the time T of the
>time stamping itself influence the results.
>
>Possible solutions:
>1. Exclude T from the result by some artificial trick. (as in the
>discussed patch)
>2. Measure T before the loop and subtract after.
>3. Make stampings once in N loops decreasing N times it's influence.
>4. ?
>
>What would you prefere ?
>




More information about the ofw mailing list