Which is more precise? TimerDiff or _Timer_Diff?

[Mithrandir]

I want to test whether using StringSplit or _StringExplode is faster but on initial tests if I use _Timer_Init and _Timer_Diff it shows _StringExplode is faster (0.17 miliseconds against 0.20 of StringSplit) but if I use TimerInit and TimerDiff it shows that StringSplit is faster. Which of the two UDFs is more accurate? Or should I take an average?

[JohnOne]

I cannot see you getting those same results every time, and .03 milliseconds it not really much to worry about.

Personally, I always use the native built in functions where I can.

[Mithrandir]

I cannot see you getting those same results every time, and .03 milliseconds it not really much to worry about.

Personally, I always use the native built in functions where I can.

Ok, actually my purpose is not the difference between stringsplit and stringexplode, what I would like to know is the difference of time between using a regular expression to extract substrings from a string or using stringsplit and then travel in the array to take the value that I want. And since with a simple comparison timerdiff and _timer_diff showed contradictory results, I didn't know what to use. Anyway I can take both and make an average I suppose.

Edited January 18, 2011 by Mithrandir

[MvGulik]

"Which is more precise? TimerDiff or _Timer_Diff?"

TimerDiff(), as it has less native autoit code overhead to contend with.

But considering TimerInit() and TimerDiff() are directly related to the CPU tick frequency ... your generally better off running a longer timed/looped test. In which case it really don't matter if you use the native TimerDiff() or the Timer.UDF _Timer_Diff().

(TimerInit(), if your really like to go nuts on the subject ... )

Edited January 18, 2011 by MvGulik