The Lag Monster Myths

I also heard the samething. I was also told that I was stack with Ginko sim. And Im wondering if thats where all my problem that i have been having is comming from.

I spoke with four or five developers about this on Wednesday. This has been tested quite a bit, as it was very important to use when moving to the class3 machines that any impact one simulator could have on the other be minimized.

The results of the tests were that pushing one simulator, even breaking it, did not impact the other simulator. My own experience matches this; many of the complaints I field are from owners whose simulator is the only one on their server; others have a "sister sim" with a much smaller load (and their sister sim has no complaints).

There are theoretical scenarios where performance can be hurt for a few minutes (mostly involving ugly crashes) but from checking real-world use, it doesn't happen just from running alongside a heavy-load sim. Whatever causes impact a simulator's performace, what's going on on the other simulator is an effective non-factor.

Thanks for the clarification! I wasn't 100% sure how the two would interact. Neither were a few folks at the SLCC!

Regards,

-Flip

The reason you haven't seen numbers for this is probably because it's subjective.

My opinion, though, is an "excessively large texture" is one that does not appropriately use its resolution versus what it is. For example, using several 1024 x 1024 textures for a simple alpha gradient would not be considered effective.





Here's a breakdown (by power of two pixels) that I tend to follow. These numbers apply to the highest value of X or Y:

Warning: This is the conservative take. Most of these are commonly upped one power of two by residents.




- 32 pixels

This is best for textures that don't need much resolution to get the job done. Very reusable for simple utilities like alpha, tint, and decals.


Examples:

* The world "Blank" texture
* Small buttons on an elevator
* Small jewelry maps
* Small alpha gradients
* Letters in very simple fonts




- 64 pixels

This is commonly a good size for textures that need a little detail, but not much. Fairly reusable.


Examples:

* A low-detail, tilable surface texture
* Larger alpha gradients
* A small image in a public area
* A small, dual-frame texture animation
* Letters in most fonts
* Most particles




- 128 pixels

This size is acceptable for most simple texture files. Not commonly reused.


Examples:

* Entire words or phrases
* Tilable textures on medium-sized objects (~1 - 3m)
* Complex particles
* Simple multiframe texture animations



- 256 pixels

This size is acceptable for most textures. Reuse is marginal.


Examples:

* Signs and a few sentences of text
* Most texture animations
* Tilable textures on larger builds (~3 - 8m)
* Reusable textures
* Very complex particles




- 512 pixels

This is the most common texture size in Second Life, which probably accounts for the lag. If you use this texture size, make sure it can be used alone with very few additional textures.

Examples:

* The world default wooden texture
* Nearly everything built at a very large scale
* Photo-realistic textures
* Advanced texture animations




- 1024 pixels

The granddaddy of all texture sizes, this is typically the largest size used. Use very rarely.


Examples:

* Works of art
* Highly-detailed and photorealistic surfaces
* Extremely complex texture animations
* Very, very large builds
* "Baked" surface elements

LOL... 4 bucks. Man, times have absolutely changed.

Yeah, I have come to the same conclusion. If we're tired of the royal runaround... best way to handle it is to stop paying LL $7,000+ a year and find something more beneficial to occupy our time.

Jeffrey, best post I've seen in a long time.

Only point we've really been making here is that when we've done about all we can client side, then it comes down to the server company providing reasonable service for our $195 a month. (And frankly folks.. that is a HUNK of money for a GAME).

When we cut scripts, when we remove tortured toruses, when we remove physics items from the sim, when we trim and use smart building techniques.. and the sim STILL lags.. but it doesn't lag all the time... just at odd intervals... and we call a Linden for help and we're told time and time again it's CONTENT when we know good and well it's not... that's when we get ticked.

We don't pay LL $2,340 a year to make lame excuses. Yeah, they have a big job. They charge big fees too. So stop making the excuses, fix what's causing the lag, get on with business.

Hey, I'm all happy happy happy that 1.7 is going to solve a bunch of problems. Goody. But I'm not at all happy that we spent weeks trying to find a lag problem and then later found out that we were blatantly lied to by some Lindens who were withholding information essential to our locating sim problems.

We don't pay that kind of money to have information withheld and facts hidden. I'm not even going to get into that here... that's coming in another post. But when we determine that the cause of excessive sim lag is NOT content and LL tells us it is... that's when we lock horns. Because we're not amateurs and we're not idiots and we know how to analyze stats.... as limited as those stat tools may be.

Lee, I 100% do not wish to be insulting at all. But I honestly can't tell whether you really mean this... or whether this is just more LL smokescreen.

Yeah, scripts CAN have variable results. By absolute definition, they are not static. That doesn't mean they are so UNSTATIC that they'll cause a sim to lag to molasses every 3-4 weeks for days on end.

Any competent programmer can analyze sim performance and come to logical conclusions. But at this point, I'm tired of discussing this. We've provided ample evidence of serious server and data flow problems on SL, PROOF of such is historical fact, and like someone pointed out prior... no matter how much data or evidence or fact is available, some people are going to continue to ignore it.

At that point, it comes down to the decision... do we keep paying LL or not? Apparently many, many users no longer are. And I tell ya, the question is on our minds every day now.... and especially every month when LL sends us a $600 bill.

Actually, yeah. We had a Linden who was actually trying to be helpful make that suggestion to us once (can't for the life of me remember who). So while the Linden was standing there... we did. Flipped the no script button.

And as astounding as it may sound, nothing happened. Sims speed remained pretty much the same. Linden says, "huh... I'll check further into that". And like so many other times, that's the last we heard of it.

That's when we started ignoring the script fairy tale. Yeah, badly behaved scripts CAN lag a sim, absolutely. Doesn't mean that everyday scripts DO. Sure, 500 scripts is going to have more effect on a sim than 5 scripts. That's obvious. But whether that 500 is going to significantly lag a script to bring it to a standstill... nah, not if they're normal, everyday, well-behaved scripts. (and please, I really don't want to hear from the noids who are about to say, "what do you consider "normal, everyday, well-behaved scripts"? If you can't figure that out on your own, you're in sorry shape.)

Again to clarify... ALL content leads to sim speed degradation. We are not speaking of standard, acceptable speed loss due to standard, everyday content. We are speaking of sims coming to serious speed drop for days and weeks on end and LL telling us "it's your content" when that same content was running just fine the weeks before. And we're tired of word-jousting with clowns who think we're such morons that we didn't think to factor possible content change into our analysis. (not you, mind you. You just made a valid suggestion.. cut scripts and see what happens).

Anyway, we did, nothing happened, someone apparently couldn't come up with a reason that didn't scream server problems, so we never heard from anyone on the matter.

Lee, sorry to disagree, but the above statement completely disagrees with our findings. I don't buy it.

We're going to be doing a major forum posting soon regarding exactly this subject. We had a dozen people assisting us with our research. So I hope you and LL are ready to bring your data to the board, because you can be sure we will be. And we're going to be revealing some facts that richly deserve to be revealed.

Flipper, don't be too quick to buy such claims wholesale. You have to remember that we were told, directly by LL employees, that sims were never doubled up on servers. So don't swallow this nonsense. 'taint true. You have to remember that no one but LL has access to testing these statements, and LL has a lot to lose by having a qualified tech prove that adjoining sims can seriously lag one another.

After being *allegedly* intentionally deceived regarding these matters for MONTHS... our patience and tendency to grant "benefit of the doubt" is at an end. Our group paid almost $3,000 to be placed on what we were told were private sims with dedicated servers. That has proved to be untrue. And when I asked LL to provide me stats on those servers, including manufacturer, model number, configuration and price of a single server, was told such would be provided-- and then it wasn't-- at this point I'm not accepting anything LL says at face value.

Can't wait for the data, I am sure many people would love to see your results.

I also thought of something else. I am not sure if I can clearly say what I am trying to say, but I will try anyway. If there are a bunch of scripts, prims, active objects, and such on a sim, and the sim can't handle it, is it a sim issue, or is it the content that is the problem? The really issue seems to be the content running on the server, so is it really possible to say it is one or the other? There are "fixes" to some of these issues, but it really seems to me that these fixes are just to not allow content to go out of wack.

Correct me if I am wrong, but does clicking the "No scripts" button only stop outside scripts from being run on the land? Not knowing what data you have collected, but I would be interested to know, if you cleared the entire sim, does the lag stay? That to me would seem to be a definitive way to see if it is content or back-end related. I know, this would mean lots of lost work, so I really don't expect anyone to do this.

1.7 is just around the corner, maybe wait and see if it helps any? I've seen really good improvements in responsiveness in preview.

Actually Dnate, that is a pretty valid question. Here's my take on that:

If the sim runs constantly slow, all the time, it might be reasonably argued that it's content.

But when the sim goes up and down, running fine sometimes and then running lousy other times, when the content stays the same... that points to server issues.

Point is... if content would be just fine if the server was running correctly, then content isn't to blame. The argument was made earlier that whatever rate the servers are running at, we just need to accept that and work around it. However, that's not what we're paying the big bucks to do. We're paying for working servers and working code. Otherwise it's just one big experiment and LL should cut its prices to beta test rates.

Content can only be cut so far before we start getting to "boy, this is a boring sim" status. My experience is you can't cut content enough to make a sim run well on a steady basis and still keep a sim interesting. If it's just an activity sim, shoot, build one large house, crank the music, invite friends and forget doing anything else with it. But if you want a beautiful sim that people revisit time and again... a certain amount of content and scripting is required.

When LL servers run that content just fine one day and then not fine the next day and that goes back and forth and back and forth, it's apparent that there are server issues-- as has been documented repeatedy on these forums by parties other than us. They've had streaming inventory problems, stacked server problems, data flow problems, etc. Now no one really is ticked about that. What folks are ticked about is being bamboozled about it. It's the old Nixon syncrome: if someone would just have the gnards enough to stand up and say, "yeah, we're messed up, here's what's going on"... there might be a lot less outrage than when people find out they've been deceived as to what's going on. Because when LL claim there is nothing going on server side, that it's ALL CONTENT.. then people start wasting their time trying to find the truth behind the matter. And when they find out that LL knew the truth all along and was withholding vital information-- that really ticks people off.

That's a reasonable statement. LL has been busting their buns to get 1.7 out and we're all awaiting the results.

But I think someone posted earlier an equally valid concern: what excuse is LL going to come up with when things in 1.7 are screwed up (as they surely will be. We don't expect it to be bug free). Since they have not been truthful with us in serious problems we've experienced in 1.6... why should we expect them to be any more forthright with us once 1.7 hits the stands and we start discovering inconsistencies in that version?

A system can only progress properly when both the hosts and clients work co-operatively with one another. That's nearly impossible to do whan a) we are not provided the tools we need to evaluate our lands and b) the host provides us with bogus information that skews our observations.

Which brings us down to the bottom line that we're tired of busting our buns and wasting our time trying to help LL improve the system when we're treated like mushrooms (and you know how folks grow mushrooms... keep 'em in the dark and pile on the manure).

I assume you still have hope for LL changing, since if you dont believe they will get better, you'd have left. So how would you want LL to change? What steps must they take to staunch your anger?

I wouldn't presume to speak for Wayfinder, but I think one can safely conclude that he'd be less irritated if LL was willing to treat what are effectively their beta customers (i.e. current customers) with the honesty that software firms traditionally do.

I know I would.

Woots! Eggy called me insane! I'm famous! Yeah, I was ranting a bit in that post. I DO love coffee and conspiracy theories. Both in quantity. Wells Fargo aside, I still think estate owners deserve better tools and support. I stand by that.

Eggy, on the issue of Wells, can you explain what you meant by "there is no Wells Fargo?" Are they totally uninvolved in Stage Coach Island? Did you just decide to set up a Wells Fargo island all on your own, because they're such a cool bank? I can dig that. But do you think they might be upset if they find out you're using their name, trademarks and reputation without even askin?

I admit they're WAY cooler than Bank of America. But an island shrine? Whoa!

I wonder if BOA would give me free checking if I set up a BOA island...

I really hope so but the preview doent have 70,000 peops in it

I'm not a techy, those statistics are chinese to me but...

I'm a customer for more then a year now and i play SL on a computer build by a techy and on cable internet: running Everquest 2 smoothly wich definatly means something (my videocard has 256 mb and my pc has 1536 ram), when most people cant move on an event i still can but...

Things have changed in a negative way in SL, and yes, all i can say as a complete non-tech: its LAG!

No lag in Everquest, no lag in Everquest 2, no lag in World of Warcraft, but shitloads of lag in SL, so its not me and my pc

So Lee, here comment from a 'average' user

What is "lag"? Reading back you will see that lag is a very broad term that is a lable for many issues. Could you define what the "lag" that you see is? Just because one game can run fine on a PC, that does not mean everythingis ok for all other games. It could be something as simple as a needed upgrade for a driver.

Dnate, someone mentioned this way back there. I'll offer two points if I may, points that have been stated before:

1) I think it's pretty much apparent to most people what the average user means by the term "lag". Folks can argue all day long about how many angels can dance on the head of a pin. When someone says to me, "Wow system sure is laggy today"... I think pretty much anyone immediately understands what that user means.

2) A lot of times, I see things mentioned in this forum that are so extremely elementary in concept that I have to wonder why folks think they're valid to even mention.

Yeah, agreed, "lag" could be caused by many things, including a person's biological state suddenly speeding up so that the entire universe around them seems to be operating at a slower level.

That's where plain old common sense comes into play. When three users are walking through a sim and suddenly all agree that "the system is laggy" today and that they've all three been having trouble with lag all day long... I don't think that's really a matter up for debate. They're paying to use the system, the system is not performing to their expectations.

I don't care what tech figures someone spouts, what the "cause" of that lag is, the bottom line is that when the end customer is not receiving the value they expect out of a product, the company is going to lose end customers. That's what it all comes down to.

At that point, it is not the responsibility of the client to try to figure out WHY the company is not delivering the expected product. It's the company's responsibility to live up to reasonable expectations-- and frankly, even more than that if their customer is important to them. That's when ya stop making excuses, stop being buttheads, and fix the problem-- no matter what that problem might be. And you dignify the client in the process by listening and responding to complaints.

Why? Because the client signs your paycheck. Without that paycheck.. you're not going to be able to offer the client anything. So that's when the company listens to the client, no matter what the client is saying, and at least makes an effort to treat the client with respect.

That's basic business 101... and it supercedes by thousands of years computer tech 101. And there's a fact of life today that people in the tech field need to realize: there are other people out here who are ALSO in the tech field, who are able to perceive situations, factor in ALL the available data, and come to logical, reasonable conclusions. Because frankly, there is a tendency among computer techs (often due both to immaturity and inexperience) to be conceited twits who think they know more than anyone else on the planet. They are insulting to other people and are condescending and rude and when I'm involved with such people, I explain to them the need to change that attitude or I hand them their final paycheck, because such people are worthless when it comes to getting down to the root of real problems. That conceit often causes tunnel-vision, and tunnel vision is the surest way to overlooking simple answers to major problems.

A lot of the posts I've seen here have to do with very, very elementary concepts. Are your graphics drivers updated? (duh) Were there the same number of avatars on the sim? (duh) Are you sure your content remained the same? (duh). Maybe there was one script that was messing with your sim. (duh) What is lag anyway? (LOL)

Like, we never thought of all those things? We never took any steps to eliminate those as viable possibilities? Because we're total morons incapable of understanding the intricacies of computer system operation?

And folks wonder why we get exasperated. LOL

Long answer to a short question, I know, but it's been said before and apparently ignored... so thought I'd say it more specifically. As amazing as it seems, some sim owners actually do know how to install video drivers. We know that other stuff too. And we do grow tired of fielding elementary-school computer concepts over and over and over.

When this thread started, we were questioning the cause of lag. We knew it was server side; we just didn't have any evidence as to what. That's no longer the case. We've had some 5-6 months to study the issue. We now know what causes excessive lag. We're not debating any more... we're publishing findings.

Now the same people who keep bringing up this computer 101 irrelevancy also seem to continue to ignore absolute statistical data. I can't do anything about such people. There are people to this day who swear man never walked on the moon. Takes all kinds.

LOL. When I read this I actually chuckled. Basic court psychology, isn't it? Turn the focus of attention on the victim rather than the perpetrator?

This doesn't have to do with my anger. That is not the focus of this discussion nor is it the issue. Staunching that anger is not the issue. Nice try Flyingroc... but let's focus on the real core issue, ok?

The question to ask is: why am I and other sim owners allegedly angry at this point? And if we are indeed angry... what is Linden Lab doing about angry clients? Now those questions I can answer:

1) People dislike paying a great deal of money for things that do not work
2) People dislike asking questions and being ignored or deceived
3) People dislike being disrespected
4) When such things happen on a regular basis, people become frustrated.
5) Continued frustration leads to anger.

What can Linden Lab do about this?

1) Start being honest with clients. Answer questions directly, accurately, and stop hiding behind BS smoke screens. Stop telling us things that ARE NOT TRUE. Basic human nature: people hate being lied to, and they equally hate asking a question and having that question totally ignored. Do we really have to spell that out?
2) We do appreciate that v 1.7 is coming out and that it will solve some of the problems being discussed. But no release date is set and the latest announcement is "it's weeks away". (Weeks? How many is weeks. 2? 4? 12?) We appreciate Linden Lab is working its tail off on that release. *We* are working our tails off trying to keep our sims running so we can pay Linden Lab bills. We think that needs some consideration as well.
3) In the meantime, people are still using 1.6 and when something goes seriously wrong, we need it fixed now... because we're paying the bills now. And when our sim lags to a standstill, we are not going to wait until 1.7 to have it repaired.
4) When we find something that's really serious (such as the prim calculation bug that destroys builds every day and is potentially a sim killer)... fix it. Now. Don't leave such a potentially dangerous bug in the system without correcting it.

That's pretty much it. It all comes down to the simple concept that if LL doesn't want angry clients... they need to start treating those clients with respect. They can work on 1.7 24 hours a day and that's not going to mean anything to a client who can't do a simple 2-prim link or whose sim is lagging so badly that their visitors can't move.

Another thing which comes to mind: when you take $1,200 from a client and tell them their sim is going to be on a dedicated server... it durn well better be on a dedicated server. Putting them on an existing server with another sim might be considered breach of contract, no matter what your tech-explanations for doing so. If you don't inform the customer and the customer doesn't have a chance to authorize or reject such a move-- you've crossed the line. That has nothing to do with tech; it has to do with basic business law.

There's lots more I could mention... but that's a start.

Thank you Leyla. I very much appreciate the common-sense simplicity of your statements. Because that's what it really comes down to.

New user visits Second Life. New user lags. New user says, "Well this sucks." New user goes decides to play Guild Wars instead. New user tells friends how badly SL sucks. They don't even bother to visit-- or if they do, it's with a negative preconception. Domino effect.

Those users don't care about LL excuses. They don't care about tech babble. They don't care about what LL is going to do in 1.7 to fix these things. They observe, they judge, they act.

This goes true for existing users too. Perhaps to less an extent, but it happens. They know how SL works. They see the operation on a daily basis, several hours a day, all over SL. They get frustrated. They lose patience. They tell SL to kiss off and they go do something else. Happens all the time. It doesn't matter whether that client knows ANYTHING about computers or tech. If they pack up and go home... their patronage and money goes home with them. Just that simple.

Old business addage: excuses lose clients. Lies lose friends.

Hate to tell you, but a time has been set. Here is the post that tells you that 1.7 is coming out the week of the 16th.

First of all, I am not sure how much programming you have done before, but to just make things work right now is nearly impossible when it comes to software. I would guess that some of these bugs are known, but they have no idea even where to look to try and correct them. Some of the bugs may be so deep into the code, that it would be near impossible to do a instant fix without breaking everything else.

I hate to keep bring this up, but look at companies like Apple and Mircosoft. Look at all the issues they had early on. Windows 1.0 was hardly even useable, yet people paid for it because of what they could get out of it was useful. Now, 20 years later they have Windows XP, which is so much better then anything that come before it. Yet, XP still has bugs and loopholes that allow things to go really wrong. Apple, I think, is even worse. They make you pay of updates to their software. You can even go back in the lifespan of SL and see how much of an improvement has been done with system stability. When I first came here, it was much worse then it is now. Some issues have been fixed, other are still around. I believe in LL and the work they do, that is why I am still staying around. Will we ever have a perfect platform? I don't think so, but as things come up, new fixes will be applied.