Script Time Awareness!

[Aryte]

Randomly, I decided to start measuring how much script time people generate (rez a cube with a basic script, check on estate menu). It's revealed some pretty interesting facts.

First off, Ordo gear is screened-- we purposely try hard to keep impact low. The SCAR-L generates .016 MS: that's great. The average, fully equipped Ordo member should not generate more than .3 MS in script time. Ruin Nefarious, with a weapon, jetpack, HUD, and cigarette, sits at .289.

The major problem is third party stuff. This could be /anything/. Color change scripts, anything with a listener, ear and tail twitches, Mystitools (do not use, omg), third party HUDs, extra sensors of any sort . . all add huge amounts of latency. So please, please be aware. Delete your tail twitch scripts, delete your ear wiggles and touches. Check your particle emitters (cigarettes), check your armor, remove any of your random walk sounds, typing override images, 3rd party HUDS, detection scripts, etc.

Compounded, high script time avatars really beat the crap out of a sim. To compare, our sim has 3.0 MS of script time that belongs to it. With 20 avatars in it, right now, the sim is at 14MS script time. That means 11MS belongs to avatars. .55MS average per. In reality, only 6 MS should belong to avatars at that number.

It takes everyone's participation! So be sure to check your avatar and remove un-needed stuff!

[Sekonda Huet]

NOT MY TAIL!! DX

[RazorFox]

I would like to know the script time in the following things:

-Kani head eye-color changer

-Aventity fox ear-twitch

-Aventity fox tail wag

[Cyphre Iredell]

I would love to hear the same, especially if Ruin is running that kind of script time. I believe on myself, with those limited attachments, plus some extras, I only hit around .2 and below (unless I am forgetting). Furries are not scriptless!

[Aryte]

You can have twitchy ears and tails off sim and not during combat.

[Syl Kiranov]

In the next SL server patch, avatars will show up on the top scripts list, so it will be easy to find the culprits. Please check your gear and sort it out before then.

Basically you should be wearing as little scripted gear as possible. Most products in SL are made by people who have no idea how to script, so don't be surprised if something small like a cigarette or a pair of boots have a higher script time than our limit all by themselves.

As the Imperator said, multitools, scanners, and combat HUDs should not be used, and here is why: Mystitool (for just one example) has 0.1 script time when PASSIVE, and its script time only goes up from there. Ordo gear should cover all your combat needs, and if it doesn't then we need to make low-lag alternatives.

[Lestat Umarov]

Furthermore, if anyone wants to see what thier script time is on sim just shoot me an IM and I'll setup a prim you can sit on and we can work from there.

[Sera Otoro]

this would be a good opportunity to check weapons in our armory... ill be poking you later lestat to see how the energy weapons are...

[Anlysia Gregoire]

As an example...I have a click-script in my helmet, a frosty breath particle thingy made by Niiya in my gas mask, a twitching tail, blinking Kani eyes, Demo's new MP5N, my AO, and my Ordo HUD that are scripted and my script time is only about... .154 I think, slightly over half the limit.

On the other hand, I have a sword with two colour change scripts and three particle prims in it, and it sits at .296. By itself.

So check your gear.

[Sekonda Huet]

I am in the process of inspecting Terra to ensure this is done. Zero's been my first victim. 83

I managed to pull nearly 16 scripts out of my avatar that where utterly useless.

[Trevor Russell]

I would like to point out that Aryte gave pretty general recommendations, not sweeping mandates. I really hope this does not turn into another fascist crusade by overzealous officers, like what nearly happened with Avatar Rendering Cost. I understand performance, but we must also be reasonable. If we want to start removing blink scripts from kani eyes, then we should certainly ban AOs, as they are much more laggy. I am not suggesting that we do so, I am simply saying, all in moderation here.

[Kellervo]

I have two weapons, an AO (an old, outdated version I bought almost two years ago no less!), and the Ordo HUD on at all times in Titan, and I was just barely scraping over .2 with a simple Kani eyeblink on top of it. It shouldn't be a serious problem trying to fit under that limit.

[Silver Meridoc]

I find this very odd in fact, having tested 3 of my most common avvies with this

My wolf avvy, primary body and biggest avvy i have, fully kit with Heavy armor, Jet Pack, P-99, SCAR, AO, and even a Tactical helment with radio script was no more then like .2-.3

Yet my Lucario, which has zero scripts, kitted out with just chest armor, shoulder, tactical belt, defro radio, AO, COM (SL chat animation) and the Fairlight Viper was .6

Lastly, Tanya, my Mutation Industries red fox, kitted only in the upper portions of Heavy armor, jetpack, fairlight pistol and TI P-90 was around .5

Apparently size isn't everything, im still hoping to clean some of those two up

[Kellervo]

Basic prims have no effect on script time. The defro, COM, and AO are probably the main reasons behind the Lucario av lagging. The TI P-90 was likely the main source of Tanya's lag. The thing's scripts are, as Tsume would put it, epicly out of date.

[Syl Kiranov]

You can examine the script time of objects by themselves to figure out where your script time is coming from too.

The following are some approximate script times. (averaged where possible, just to give a basic idea... may change under some conditions.)

Ordo SCAR-L: 0.016 average when drawn (not including sling)

Ordo utility HUD: 0.06 average (it's up and down a lot, had to tally for an average)

Ordo Defero unit: 0.005 average (with no com chatter)

Fairlight pistol: 0.05 average for one pistol when drawn. (Holster averages 0.008)

Fairlight viper: 0.05 average each when drawn. (Has no holsters.)

Titan M468: 0.05 average (I don't have the P90 so maybe this is similar?)

Zhao2 (idle): 0.014 average when standing. (much higher when moving.)

Keep in mind some things will have big script time spikes under certain conditions too, for example anything with a channel 0 listen.

[Lestat Umarov]

I managed to get myself down to 0.020 without the HUD and a weapon. If I can do that with my furry avatar there's no way you can't do just as good or better.

[Trevor Russell]

I went from .625 to .3, wearing a jetpack, new armor, SCAR-L (slung), and utility hud for ordo stuff.

Switching to ZHAO-II lowered it a bit.

Detaching the Kani HUD was huge, along with Kristmas Kitty HUD (which was only a tiny part of the change) i dropped .2

(Kani HUD has a color bar, made before touch sensitivity, so there are 30-40 increments that each have a script @_@)

Detaching a Multitool (custom by Faron) i dropped another .05.

Removed tail scripts for .025.

Removed kani eye blinks, tongue sticking out, and eyebrow movements for .05.

If you, like i did, think "BUT BUT, I LIKE TO EXPRESS MY FURRINESS!" then ask yourself, do you really need blinking eyes? will anyone notice? the answer is no. And if they do notice, they probably won't care. Just because your eyes blink will not make the difference over whether or not you get laid that night. Instead, think of how much of an improvement to combat can be had with less lag!

[Cyphre Iredell]

Those silly blinking eyes spam channel messages too :/

[Esichs Sonnerstein]

Not to mention, if Uchi did his own scripting, it's gogin to be epicly lagy regardless. Guy can build, but cant script to save his life D:

[Kellervo]

Well, the eyeblink replacement script that Faril has is -much- more efficient then the original script, and can be safely used without even wearing the Kani HUD. It also adds more colors. 'sides, the only reason you'd need the Kani HUD is if you're constantly changing emotions / ear positions, as the eyes themselves can be controlled via the /1 listener, so really, it's something that shouldn't be constantly worn.

That right there, for half the Ordo, is an easy .2 - .15 off. Last I checked I hovered around .175 WITH the eyeblink still intact, albeit with only one weapon instead of the XAR9 + P99.

[Gulliver Carpool]

I'm bumping this.

As I understand it, ARC has nothing to do with script time. So, things like wearing a slung SCAR shouldn't be verboten correct?

[Syl Kiranov]

Correct, ARC is totally unrelated to script time. ARC is there purely to give people a basic idea of their visual processing cost (what they're doing to people's framerates.) We don't have a rule on ARC that I know of, and it's kind of a broken function (the results are extremely inaccurate) but you probably want to voluntarily keep it below a huge number like 10,000.

The slung SCAR's script time is minimal... a single script with a single touch event. It will probably bump your script time up by 0.009 or something ridiculous. I personally don't wear it due to the rendering cost, but you won't see a noticable script time difference.

edit: Forgot to mention this (and it should probably go without saying) but it's considerate to remove any extra attachments such as the SCAR sling and even the SCAR itself during weekly meetings. Small things add up when there's 30 of us in the same room.

Edited May 8, 2009 by Syl Kiranov

[Emma Roy]

bump!

[Sera Otoro]

ill be asking someone again to be checking the script times of all the ordo weapons i scripted... in the mean time im about to start the new updates, so dont check them yet...

[Kytec Switchblade]

The estate menu script times are flawed. It can give an average while idle, but there are a few factors to consider. The hud scales in latency when the av radar picks up more people. I've had my script time jump averages between 0.21 and 0.8 just by going from checking at 3000 meters to checking in a meeting. The only extra script I have is blink on my new head, which doesn't seem to generate much, and listens solely for it's hud on some huge channel in the negatives or some such. I don't know about guns creating latency, but i know 30 people standing around and 30 people firing full auto makes a world of difference, maybe we should test script time idle and script time firing full auto.