answered In Mono, is llGetListLength() actually 100% faster than (some_list != []) ?

[Love Zhaoying]

9 minutes ago, Quistess Alpha said:

I don't believe "elapsed time that was allotted to the script" is something we have access to, which is the problem. maybe someone could file a Jira to extend llScriptProfiler()?

I think that llGetTime() is close enough for me, unless I am missing something in the explanations above.  Like, if it is wildly inaccurate and/or inconsistent between loops/runs,etc.

[Wulfie Reanimator]

23 minutes ago, Love Zhaoying said:

I think that llGetTime() is close enough for me, unless I am missing something in the explanations above.  Like, if it is wildly inaccurate and/or inconsistent between loops/runs,etc.

It's not useless, but for the purpose of profiling a script's performance, it lacks resolution (being limited to server frames), and it's inflated on top of that because it doesn't track running time (time passes even if the scheduler doesn't run the script).

[Love Zhaoying]

13 minutes ago, Wulfie Reanimator said:

It's not useless, but for the purpose of profiling a script's performance, it lacks resolution (being limited to server frames), and it's inflated on top of that because it doesn't track running time (time passes even if the scheduler doesn't run the script).

I'm only using it to track the script's "relative" performance - the performance of each function in the script compared to each other function, and each "atomic" action within each function, to each other "atomic" action.  Does that help with my use-case for using llGetTime() for profiling?

[primerib1]

Well for my use case, it's just to validate the statement in the title of the thread.

And it works. Well, ballpark-ly.

But I do agree requesting a better Profiling tool...

 

[Quistess Alpha]

1 hour ago, Love Zhaoying said:

I'm only using it to track the script's "relative" performance - the performance of each function in the script compared to each other function, and each "atomic" action within each function, to each other "atomic" action.  Does that help with my use-case for using llGetTime() for profiling?

No not really. I think the theoretically best method for relative performance testing would be to run both functions simultaneously in separate scripts, for a predefined interval of time (say, 5 seconds) and have them each report the number of repetitions. That's a bit of a hassle to set up properly though.

[primerib1]

53 minutes ago, Quistess Alpha said:

No not really. I think the theoretically best method for relative performance testing would be to run both functions simultaneously in separate scripts, for a predefined interval of time (say, 5 seconds) and have them each report the number of repetitions. That's a bit of a hassle to set up properly though.

You can (mis)use LSD for syncing...

Have both script wait for a certain key to actually no longer return ""

Do an iteration.

At end of iteration check if the key has returned to ""

If yes, stop & report. If not, lather rinse repeat.

5 seconds might be too short, though. Make it, say, 30 seconds.

Edited January 16, 2023 by primerib1