Jump to content

Questions about Draw Weight


Diovona
 Share

You are about to reply to a thread that has been inactive for 3955 days.

Please take a moment to consider if this thread is worth bumping.

Recommended Posts

I have two questions about draw weight.

#1 Different draw weight values on different computers

Last week a friend and I noticed something weird about the performance tool "Show Draw Weight Weight for Avatars" in the Advanced Menu: it shows completely different values on my computer and on her computer for the same avatars. The difference was from 50k to nearly 500k.

Any idea what's up with that?

#2 How will draw weight change with the new materials viewer?

What effect will the new materials viewer and server side baking have on draw weight? That is, what will change considering the factors of an avatar's appearance that contribute to lag at bit events? So far, I've always checked my draw weight and the script count. Will that stay the same? Are there new settings coming up for us to check in terms of lag?

Link to comment
Share on other sites


Diovona wrote:

I have two questions about draw weight.

#1 Different draw weight values on different computers

Last week a friend and I noticed something weird about the performance tool "Show Draw Weight Weight for Avatars" in the Advanced Menu: it shows completely different values on my computer and on her computer for the same avatars. The difference was from 50k to nearly 500k.

Any idea what's up with that?

Is the person seeing the higher number using a V1 based viewer and the one seeing the low number a V2 based viewer? Or was it the other way around? I forgot. Anyway, when mesh was introduced, the avatar rendering cost was changed into avatar draw weight with a different calculation. The avatar draw weight should be under 40k according to LL, so if that way of calculating rhymes with the 500k one of you sees, the person with that weight is wearing some serious lagging outfit.

 


#2 How will draw weight change with the new materials viewer?

What effect will the new materials viewer and server side baking have on draw weight? That is, what will change considering the factors of an avatar's appearance that contribute to lag at bit events? So far, I've always checked my draw weight and the script count. Will that stay the same? Are there new settings coming up for us to check in terms of lag?

 

 

 

I don't think there will be a lot of change. The avatar draw weight is based on how much efford it takes your computer to draw the fully rendered avatar on your screen. In other words, mainly GPU load. SSB won't change that. Added normal maps and specular maps could increase the load slightly.

Link to comment
Share on other sites


Kwakkelde Kwak wrote:

<snip>



. The avatar draw weight should be under 40k according to LL,


</snip>


When/Where did LL ever say this?

Unless you are referring to the fact under 40k is what some <word redacted> at LL set as the green zone for draw weight.

Under 40k is essentially a naked Avatar wearing no attachments.

See "So Linden Lab Wants Us All To Be Naked?" (pops)  for a discussion of this ridiculous setting.

Link to comment
Share on other sites


Kwakkelde Kwak wrote:

I thought I read it in the wiki once, can't find it though. Anyway, the fact it turns color at 40k should be enough to interpret what (someone at) LL finds reaonable.


Yes, someone at LL thought it was a reasonable number.  Whether it was just something they pulled out of the air or not is another topic.

However, if there are decisions that they (LL) are making based on that 40k number, decisions that might effect performance, then they are out of touch with reality in SL.

You are correct that there are things that can be done in content creation to reduce impact.

Ocasionally when I am having trouble I will take a look at what my own impact is and occasionally I will get shocked.  Like finding an object I am wearing has 5000kb in scripts.

But really I think for the large majority of users (and yes I know that is just my opinion) all they want to do is log in and enjoy thier SL without having to mess with these details.  And based on my looking around as I did in the thread I linked, around 100,000 is a pretty avererage draw weight for people in SL.

Link to comment
Share on other sites


Perrie Juran wrote:


Kwakkelde Kwak wrote:

<snip>



. The avatar draw weight should be under 40k according to LL,


</snip>


When/Where did LL ever say this?

Unless you are referring to the fact under 40k is what some <word redacted> at LL set as the green zone for draw weight.

Under 40k is essentially a naked Avatar wearing no attachments.

See
  for a discussion of this ridiculous setting.

It was Nyx Linden who came up with the number of 40k. See his comment in this Jira.

https://jira.secondlife.com/browse/STORM-1679

Link to comment
Share on other sites


Perrie Juran wrote:

But really I think for the large majority of users (and yes I know that is just my opinion) all they want to do is log in and enjoy thier SL without having to mess with these details.  And based on my looking around as I did in the thread I linked, around 100,000 is a pretty avererage draw weight for people in SL.

 So, I guess I better cut down being close to 200k hehe

Seriously, despite all the debate over the years, draw weight has been severely overstated. I've never had a problem with it graphically. The numbers always seem to change. Back in the day 40k was a lot. /me looks for an old piccie :matte-motes-wink-tongue:

Link to comment
Share on other sites


arton Rotaru wrote:


Perrie Juran wrote:


Kwakkelde Kwak wrote:

<snip>



. The avatar draw weight should be under 40k according to LL,


</snip>


When/Where did LL ever say this?

Unless you are referring to the fact under 40k is what some <word redacted> at LL set as the green zone for draw weight.

Under 40k is essentially a naked Avatar wearing no attachments.

See
  for a discussion of this ridiculous setting.

It was Nyx Linden who came up with the number of 40k. See his comment in this Jira.


Some of the things I thought reading that are not printable.  Just wow!

 

Nyx Linden added a comment - 31/Oct/11 8:29 AM

"Let's go with 40,000. That seems to be a reasonable cutoff for avatars that are "reasonable". Most of the worst performance offenders are MUCH higher than this, and it will give us some color variation in the less complex avatar's displays to encourage lower numbers."

Did he even have any idea what the average Avatar weighed in at?

And the statement, "to encourage lower numbers," while he probably meant for individuals to take control of their own numbers, that could be read as encouraging another ARC War.  Talk about the 'law of unintended consequences.'

Really, how many people would take the time to go through all their outfits to make sure they were not "performance offenders."

Perhaps LL should mandate that all Avatar Appearance Providers (clothes, hair, jewelry, etc) state the draw weight for their products.  Maybe we as the customer should be demanding that this be done so we don't become inadvertent offenders.

Talk about surreal!

Link to comment
Share on other sites

Well yeah, 40k isn't that much wearing legacy content. But, obviously Nyx developed this new formula with optimized mesh content in mind. Which is the right way to go IMHO. Sure, there's still all that sculpty, twisted tori shoes, hair what ever out there, and people will still wearing these. But that doesn't mean that the formula should accommodate that very inefficient content, just because it's there.

Link to comment
Share on other sites


arton Rotaru wrote:

Well yeah, 40k isn't that much wearing legacy content. But, obviously Nyx developed this new formula with optimized mesh content in mind. Which is the right way to go IMHO. Sure, there's still all that sculpty, twisted tori shoes, hair what ever out there, and people will still wearing these. But that doesn't mean that the formula should accommodate that very inefficient content, just because it's there.

The problem with that is the "optimal," especially in SL, very rarely happens.

That would be in my defintion a "pipe dream."  ;)

Especially when LL does not have in place what it takes for the optimal to really happen.

Is reducing Avatar Weight a worthy goal?  Sure.  But you still need to get there.

Link to comment
Share on other sites

Yeah I know. But what should they do instead, providing a formula that makes no sense to account for performance reasons? Because we are all pretty bad performance offenders otherwise?

I bet that number of 40k is already pretty "generously" taken. Could have been even worse. :matte-motes-nerdy:

 

Link to comment
Share on other sites

Nobody did, I simply pointed out that we can't compare screenshots or any other records of the numbers from "back in the day" to the more recent ones.

It's just to prevent confusion. If people see that in 2006 they had a number over their head about ten times smaller than they do now, it doesn't mean they now cause more graphical load on servers, networks or people's computers.

If by "back in the day" you mean more recent, after avatar draw weight replaced avatar rendering cost, 40k was never a lot as far as I have seen. Your average prim or sculpted hair or boots have always pushed that number far beyond the green zone.

Link to comment
Share on other sites


WolfBaginski Bearsfoot wrote:

I do have one or two AVs that had a very low Avatar Rendering Cost under the old system. It was hard to get a green rating back then. I'm not sure that the current system is any better a guide.

Neither am I, but with mesh it's much easier to be "efficient" than it was with prims and sculpties. So it shouldn't be that hard to get the number on old avatar attachments down by replacing those sculpts and tori with mesh.

Putting a number on rendering cost in SL is useful. Putting a colour on it not so much.

If you always hang out in your private space with two of your closest friends and all three of you have a state of the art computers, the "green zone" could be at a much higher number than when you usually visit crowded clubs where lots of people with older hardware dwell.

I can imagine LL made an (educated??) guess on what would result in an environment where most people wouldn't encounter too much lag. That's obviously hard to measure. LL can measure what kind of impact 40 avatars with a certain ADW have on their end in a certain situation, servers and network. It's very possible that's thrown into the mix of the big guess as well. It's also possible LL decided on 40k because that rhymes with a certain amount of geometry on screen at any given time, but the way a region is set up makes as much difference as the objects in that region. If 40 people stand on 10 square feet the load on all systems will be higher than when those same people are spread evenly across the land.

All we can say is "the lower, the better". If you find two outfits that are similair and one has a weight 5 times higher than the other, I wouldn't have to think twice about which one I'd get. For builders it's a useful tool as well, although in most cases a well thought out and well built item will have a low impact anyway, the number would just confirm the fact the builder made an efford.

Link to comment
Share on other sites


Kwakkelde Kwak wrote:


WolfBaginski Bearsfoot wrote:

I do have one or two AVs that had a very low Avatar Rendering Cost under the old system. It was hard to get a green rating back then. I'm not sure that the current system is any better a guide.

Neither am I, but with mesh it's much easier to be "efficient" than it was with prims and sculpties. So it shouldn't be that hard to get the number on old avatar attachments down by replacing those sculpts and tori with mesh.

Putting a number on rendering cost in SL is useful. Putting a colour on it not so much.

If you always hang out in your private space with two of your closest friends and all three of you have a state of the art computers, the "green zone" could be at a much higher number than when you usually visit crowded clubs where lots of people with older hardware dwell.

I can imagine LL made an (educated??)
guess
on what would result in an environment where
most people
wouldn't encounter
too much lag
. That's obviously hard to measure. LL can measure what kind of impact 40 avatars with a certain ADW have on their end in a certain situation, servers and network. It's very possible that's thrown into the mix of the big guess as well. It's also possible LL decided on 40k because that rhymes with a certain amount of geometry on screen at any given time, but the way a region is set up makes as much difference as the objects in that region. If 40 people stand on 10 square feet the load on all systems will be higher than when those same people are spread evenly across the land.

All we can say is "the lower, the better". If you find two outfits that are similair and one has a weight 5 times higher than the other, I wouldn't have to think twice about which one I'd get. For builders it's a useful tool as well, although in most cases a well thought out and well built item will have a low impact anyway, the number would just confirm the fact the builder made an efford.

"Putting a number on rendering cost in SL is useful. Putting a colour on it not so much."

I do think that most of us who post here understand that rendering cost affects performance.

I also think that most of us understand that there are "best practices" when creating content to minimize impact.

What troubles me here is IF LL is making decisions that affect performance based on that number then those decisions are out of touch with the reality that very very few Ava's are in the green zone.

LL is investing a lot of resources (translated, money) in improving performance.  But the one place that they appear to be investing next to nothing in is educating people in efficient content creation.  And that is the one thing that could have the greatest long term positive impact on SL.

One thing for sure, I don't want to see the return of ARC Police to the grid.

Link to comment
Share on other sites

You are about to reply to a thread that has been inactive for 3955 days.

Please take a moment to consider if this thread is worth bumping.

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...