Measuring script execution speed

Hi -

If I make a script, say, that takes the time, counts to a million, takes the time again, and then subtracts time1 from time2, I get an output of a certain number of seconds.

What am I actually measuring here? Assuming, say, that I am floating at 3000 meters and not downloading any textures, am I seeing any kind of valid comparative measure of script execution speed?

Thanks!

Not a lot really, just the beginning and end times. Unfortunately it's notoriously hard to get a really good measure of how fast a script can get its work done, because scripts are near the bottom of the totem pole for getting a shot at the CPU, and they have to share a limited time slice with all the other scripts running on the sim.

Those won't really have a bearing on script execution. It takes some heat off what your viewer has to render, but the script still has to compete with the stuff happening on the ground or elsewhere on the sim.

the timers you see with shift-ctrl-1 are a little more helpful. If the frame rate is near 45 and time dilation near 1.0, you have a lightly loaded sim and a better shot at an accurate result. You can also expand the timers at the bottom of that panel to see how much load there is.

you are measuring how fast the code between your get time statements run on the sim you are on. in addition to what viktoria said this might help
http://wiki.secondlife.com/wiki/LSL_Script_Efficiency#How_Fast_Does_That_Code_Run

hmm that micro-optimization stuff probably needs a revisit for mono to check what matters, huh?

the whole page does really.... for instance the memory saving optimization isn't supposed to work under mono. maybe if I get bored I'll attack some of it.

Sure you can say that, but it can be hard to say exactly what it is you are measuring!

Don't know what all of the hulaballoo is about. Using script time to measure performance works just fine. Find an empty sim, such as a new island or go to Aditi and find an empty sim. TP around until you find one running around 1.0 ms of script time and then run your script multiple times to get an average reading. You can use it to see how different changes in a script make it more or less efficient. LLGetTime was what we all used to see the difference between LSL and MONO and red flagged how MONO treats state changes for example.

Keep on being curious and experimenting Oberon. It is the mark of a future great scripter when they get to the point of worrying about sim resources and how to make their scripts work better, instead of just working.