Welcome to the Second Life Forums Archive

These forums are CLOSED. Please visit the new forums HERE

i love island sims (or more research into lagging)

Samhain Broom
Registered User
Join date: 1 Aug 2004
Posts: 298
09-29-2004 12:34
I'd like to be so fortunate to have 160 prims on my 512^m chunk of real estate. Any 512^m place I have seen has always had 117 prim content.

(just my two cents)
_____________________
rm -rf /bin/ladden #beware of geeks bearing grifts
Zax Zadoq
You can't see this title.
Join date: 24 Sep 2004
Posts: 64
09-29-2004 13:55
Even better, then, for the servers. That then lets the server have just under 15,000 prims, which is less than the server can handle in scripted prims (based on others calculations).

From another thread, we've been having discussions on texture size and lag. That seems to have far more effect on lag than most things, but it would point to network lag and not processor lag. That should mean, in effect, that during times of heavy lag the server FPS may not always be very low. I know my client FPS is always much, much lower than the servers.

I do notice when I teleport into a new area and lots of data starts flowing in that the lost packet rate increases some. I wonder how many connections it's trying to make? My DSL bandwidth is 6 megabit and the client only allows up to 1 megabit.

Oh well. Ever since the days of multiplayer online games began lag has been an issue.
Alondria LeFay
Registered User
Join date: 2 May 2003
Posts: 725
10-05-2004 22:55
From: His Grace
if you create 128 prims with the "listen" script (included at the end of this post), an empty sim will go from 11,000 Sim FPS to 6000 Sim FPS. Surprisingly, this is even if they don't have anything to listen to (i.e. no one is saying anything). And even if all the objects have something to listen to, the Sim FPS only goes down by an additional 500.

if you create 128 prims with the "sit" script (included at the end of this post), an empty sim will remain at 11000 Sim FPS.

if you create 1024 empty prims, an empty sim will remain at 11000 Sim FPS.
[/i]

You know, I don't think this is a steady matrix.. To test my guess that listens are not that distructive, I rez my ancient chess board, which contains 64 listens (one per square). (Yes I know, that is ridiculous, but it was like over a year ago and people's mindsets were a little different). The sim I tested it in was Miramare. This is what I found:

Number of boards Listens FPS RunTasks
0 0 893 0.8ms
1 64 844 0.8ms
2 128 772 0.9ms
3 192 752 0.9ms
4 256 717 1.0ms
5 320 678 1.0ms

Thus, the hit is definately not as drastic on a "living" sim (or I suspect the FPS may be a bit, well, arbitrary.).

Conclution:
It seems every 64 listens the runtasks go up by approximate .05ms. The FPS drop is drastically lower than the original test by His Grace.


Details on the chess board:
Each iteration of the chess board contains 68 primes.
Each square of the board has a script in it, that among other things, has a llListen() call that is always active (on channel 240240)
There are another 5 scripts running the backend logic/control.
The backpain is running a texture based animation.

---------------------
Further testing (sorry folks in Miramare tonight)
I decided to see what would happen if I triggered these 320 listen events. Results: Small (approximately 25FPS/no runtasks) which probably would more account for the horrid else/if chain hooked to the listen() event.
----------------------
I accidently touched one of the screens, which started all the boards. What this resulted in was 325 objects changing their textures at once. Now that was a hit! FPS dropped to about 149 for a few seconds, runtasks went up to about 5.4(!!). Mind you, they quickly reverted back...
Eggy Lippmann
Wiktator
Join date: 1 May 2003
Posts: 7,939
10-06-2004 02:11
I seem to recall reading in some release notes that listeners were a little optimized recently, so your comparison to His Grace's results may not be yielding valid results.
1 2 3