What is the verdict on the latest Firestorm release, performance floater and autotune?

[Echelon Alcott]

I'm exploring the new "Improve Graphics Speed feature. Can we trust that "Avatars nearby" and "Your avatar complexity" offer better metrics that the previous ARC algorithm?

What is your experience with Autotune?

[Persephone Emerald]

I laughed when I read we'll be able to tell which avatars are lagging us the most. My next stop is the Marketplace to look for throwable tomatoes.

[Mollymews]

17 minutes ago, Echelon Alcott said:

I'm exploring the new "Improve Graphics Speed feature. Can we trust that "Avatars nearby" and "Your avatar complexity" offer better metrics that the previous ARC algorithm?

Beq Janus has posted on here in recent weeks quite a lot about what they Firestorm have been doing. I read that at the18 March TPV meeting, this subject (arctan) came up in the conversation

Quote

[13:45:23] Beq Janus: I would say that the perf floater (at least my version) is aiming to inform a future arctan by making the render costs more obvious
[13:46:25] Vir Linden: Yeah, looks very valuable for that

which is a good indication that Beq's thoughts/work is significant

 

[Myntz Mysterious]

I havent used the auto tune part yet, but have been walking around using the performance floater to see what values it puts out.

Enjoying the 360 snapshot feature!

Sadly ground sit is still broken... 😒

Haven't played with the AO either, I understand that it got an overhaul.

[Echelon Alcott]

Am I doing something wrong, or is the Improve Graphics Speed window fixed size? I wish I could resize it.

[Ayashe Ninetails]

Wait, Autotune? Like...I could run around sounding like T-Pain? That Autotune?

[Myntz Mysterious]

7 minutes ago, Echelon Alcott said:

Am I doing something wrong, or is the Improve Graphics Speed window fixed size? I wish I could resize it.

Sadly I dont think you can  I just been walking around with a big black  window on my screen...

Inara has a good write up on the new features: https://modemworld.me/2022/03/22/firestorm-6-5-3-performance-and-photos-and-more/

[Scylla Rhiadra]

1 hour ago, Persephone Emerald said:

I laughed when I read we'll be able to tell which avatars are lagging us the most. My next stop is the Marketplace to look for throwable tomatoes.

I was at a club or something once . . . what was it? I don't remember. They had a big board that listed the top "offenders" in the place in terms of running scripts.

I think I was #2 or something. SO embarrassing.

[Coffee Pancake]

Oh this will end well.

The first thing this makes clear is the ARC numbers are hot potato garbage and utterly meaningless when it comes to actual rendering cost. Will this "inspire LL" to fix up ARCTan and give us numbers that are meaningful ? Nope !

Typical Firestorm though, truncate ARC to 2 digits so it's easier to read and make space to show µs to 2 decimal places instead. A+ Gold Star.

µs is micro seconds btw, that's 1x10-6 seconds. Or a millionth of a second. A wonderful unit that I'm sure everyone understands and remembers from kindergarten.

 

They are very bad, make poor life choices and are personally responsible for your poopy frame rate .. and your daddy ignoring you.

 

Which begs the question, who decides how many µs is the right amount, because there has to be a right amount, everyone know this, just like everyone has a solid and reasonable grasp of exactly what a µs is.

Will it matter than your µs and my µs will be different and dependent on our specific setups, nope ! Do you think anyone will get that distinction .. nope !

Did you see that guy with over 9000 though ... daaam

 

 

 

Naming and shaming is not an answer. 

No one set out to be the singular cause of everyone else having a miserable time, but that's what happens when you hang a lantern on individuals. It's not going to be "that person's avatar is technically the slowest to render in µs", it's going to be singling people out as the cause of all that's not quite meeting expectations .. which as this is SL is going to be a long list.

Is your club quiet today? Is no one talking in local chat? You haven't had an IM in hours? The marketplace is so slow .. ITS THAT PERSONS FAULT ! 

Is there lag? Lets not look for the actual cause, lets just start harassing the people with the biggest numbers, Firestorm say's it's their fault.

Will this kick back to the creators of the items that cause firestorm to single people out .. hell no! 

 

 

 

The rest of autotune is probably fine and a wonderful job.

 

1 hour ago, Scylla Rhiadra said:

I was at a club or something once . . . what was it? I don't remember. They had a big board that listed the top "offenders" in the place in terms of running scripts.

I think I was #2 or something. SO embarrassing.

The fun part is .. those script counter things are pure snake oil. They are not an accurate indication of actual memory usage, only the worst possible case which is extremely unlikely. It would be as just accurate as blaming brunettes for the lag. But we all know how those things were wired into security to auto eject people, because it's not SL unless we can identify exactly who is to blame and publicly shame and punish them.

Edited March 22, 2022 by Coffee Pancake

[Wulfie Reanimator]

I appreciate the raw "how much time does it take to render X" as a useful metric, because it directly tells you why your FPS is dropping, but I wish it was more specific instead of (seemingly) being limited to just people's avatars as a whole. I'd like a breakdown by individual attachment, and an addition to the Inspect Object window for rezzed objects.

Coffee is right that people absolutely can't behave themselves with a feature like this, but...

 

Edited March 22, 2022 by Wulfie Reanimator

[Rowan Amore]

In a busy spot last night, I tried opening the sidebar for avatars and crashed so...yeah. ..wasn't helpful in that respect.

 

 

[Lewis Luminos]

So this is interesting - my two girl alts both use Belleza Freya body primarily, and Erika by preference (although less often due to the smaller amount of clothing available).

Now I was always under the impression that Belleza bodies, while nice looking, are a poorly-optimised hot mess.  So I did a little testing, checking the avatar status whilst wearing only the body and head, no other attachments or huds at all.  

With Erika, it hovered around 1200µs. With Freya it was 900µs.  Once I added back all the other things, hair, clothes, shoes etc, there was no significant difference between the two, at around 1400-1500µs.

I like this tool, I think it will be useful. 

[Echelon Alcott]

I appreciate the spirit of the Improve Graphics Speed functionality, and the attempt to provide us render time in microseconds as an alternate metric besides ARC to determine avatar cost in terms of rendering and viewing, so we can put together avatars that cause less lag.

Having played with it for a bit, and gone to a couple of sims populated with a number of avatars, I noticed that the time to render an avatar is proportional to the distance to my camera. An avatar that had very high render time when my camera was close to him/her, dropped in the list when I moved my camera away, closer to another avatar in another part of the sim. It seems like in order to fairly measure avatar speed time with this tool, all avatars in question should be equally distant from the camera. If my observations are correct, it reduces the practical use of this functionality, and might mislead people if they don't truly understand how this tool works.

Kudos on the work done. Now people just need to educate themselves and understand how it works, to use it fairly and correctly in its current form.

Edited March 22, 2022 by Echelon Alcott

[Lindal Kidd]

I don't see it as much different, socially, from Avatar Health > Show Avatar Complexity. That one shows everyone's Complexity figure above their heads, rather than in an ordered list, but it still "hangs a lantern" on high complexity avatars.

The question is, are these numbers more meaningful than the Complexity number? And, how many arguments are we going to have about which figure "really" equates to resource use?

[panterapolnocy]

35 minutes ago, Lindal Kidd said:

The question is, are these numbers more meaningful than the Complexity number? And, how many arguments are we going to have about which figure "really" equates to resource use?

You need to decide for yourself if they're more or less meaningful. Complexity is a general value reported to the simulator by all viewers present in the vicinity, calculated based on a fixed formula. Performance floater is measuring how hard it is for your computer to render everything around - that's why distance, LOD and graphic settings matter - and the results will generally differ between two different (more/less powerful, different settings etc.) machines. See:

https://beqsother.blogspot.com/2021/09/find-me-some-body-to-lovebenchmarking.html
https://beqsother.blogspot.com/2021/11/i-dont-wanna-mesh-with-nobody.html
https://beqsother.blogspot.com/2021/12/upgraders-of-lost-arc.html
https://beqsother.blogspot.com/2022/03/how-to-use-new-firestorm-performance.html

Edited March 22, 2022 by panterapolnocy

[s2Pandora]

From the bodies i've tried, Legacy jumps me the highest (easily over 9k).

Complexity is still a much easier number to read, but this is still amusing none the less.

[Coffee Pancake]

55 minutes ago, Lindal Kidd said:

I don't see it as much different, socially, from Avatar Health > Show Avatar Complexity. That one shows everyone's Complexity figure above their heads

Your own complexity number is generated by your own viewer. Some viewers fudge the number submitted to SL, for reasons.

People didn't end up doing anything socially with ARC after it was quickly found to be hot potato garbage, there was an early attempt to police it and eject people from places .. which lasted just as long as it took moderators to change outfits.

 

The µs number, that's actual process time, that's a real number (to 2 decimal places !!)

... which has to be taken in the wider context of what the viewer is doing at the moment, the overall frame time, general graphics card load, capabilities of the computer in question and so on ... the order of the avatars might not be the same for everyone all the time, and could end up very different on different viewers as we all try to reign the avatar processing time .. but hey .. THAT GUY, If we eject that guy everyone's frame rate goes up!!

 

It's a very good, very technical number, with zero context, but if you're over 9000, well, that's actually you.

[Beq Janus]

14 hours ago, Echelon Alcott said:

I'm exploring the new "Improve Graphics Speed feature. Can we trust that "Avatars nearby" and "Your avatar complexity" offer better metrics that the previous ARC algorithm?

What is your experience with Autotune?

Can we trust the numbers. That depends on what you mean. The numbers are the amount of time (averaged over the last few frames to stop it being super jittery) that that avatar took to render. It is an accurate reflection of the CPU cost on your machine. So in that sense it is a far more reliable stat than the utter nonsense of ARC, However, do not take my word for it, you have the tools now to do a little home science experiment. go to a busy place and look at this screen, you'll find some people are super high cost and others are not. if you derender a high cost individual then that amount of time will be removed from the total frame time and your FPS will increase by the same proportion. If you lookdown the list you'll probably find a High ARC person with a low render time (now this may simply be that they are out of view, but it is not necessarily so) ideally find a high ARC, low render time person in full view. derender them. Did your FPS zoom up in line with the ARC or inch up as suggested by render time? 

A lot of this hinges around the high impact of alpha cut mesh bodies. My blogs such as https://beqsother.blogspot.com/2021/12/upgraders-of-lost-arc.html show the issue, with some of the earlier ones going a lot deeper into the technical issues. 

12 hours ago, Echelon Alcott said:

Am I doing something wrong, or is the Improve Graphics Speed window fixed size? I wish I could resize it.

It is indeed. I tried to make it resize nicely and the UI system bit my arse and puked up all over my homework. I tried, honest.

10 hours ago, Coffee Pancake said:

The first thing this makes clear is the ARC numbers are hot potato garbage and utterly meaningless when it comes to actual rendering cost. Will this "inspire LL" to fix up ARCTan and give us numbers that are meaningful ? Nope !

Typical Firestorm though, truncate ARC to 2 digits so it's easier to read and make space to show µs to 2 decimal places instead. A+ Gold Star.

µs is micro seconds btw, that's 1x10-6 seconds. Or a millionth of a second. A wonderful unit that I'm sure everyone understands and remembers from kindergarten.

I did not actually do the ARC truncation. This floater revived the short-lived performance floater that LL released last June. It was the fact it centred on the misinformation of ARC that riled me up enough to do this. That floater truncated the numbers.

as for usecs..yeah I pondered long and hard about this. Ultimately I cannot convert it to a more practical unit that makes it sensible. In disclosing the actual numbers I am hopefully giving those who are interested a new tool.

This changes nothing in terms of "shaming", it at least stops pointing the finger at the wrong people like ARC does. As Wulfie says, it doesn;t really matter what we do, someone will use it for reasons other than self improvement.

2 hours ago, Wulfie Reanimator said:

I appreciate the raw "how much time does it take to render X" as a useful metric, because it directly tells you why your FPS is dropping, but I wish it was more specific instead of (seemingly) being limited to just people's avatars as a whole. I'd like a breakdown by individual attachment, and an addition to the Inspect Object window for rezzed objects.

I had that, however, due to the way that SL works it is not possible to retrieve the another avatars attachments by name without selecting them (this is how the "inspect" works) doing this would be highly disruptive and counter to what we are trying to do so I had a choice of leaving the attachments as UUIDs and thus pretty meaningless or simply removing them and focussing on our own. I chose the latter.

2 hours ago, Lewis Luminos said:

So this is interesting - my two girl alts both use Belleza Freya body primarily, and Erika by preference (although less often due to the smaller amount of clothing available).

Now I was always under the impression that Belleza bodies, while nice looking, are a poorly-optimised hot mess.  So I did a little testing, checking the avatar status whilst wearing only the body and head, no other attachments or huds at all.  

With Erika, it hovered around 1200µs. With Freya it was 900µs.  Once I added back all the other things, hair, clothes, shoes etc, there was no significant difference between the two, at around 1400-1500µs.

I like this tool, I think it will be useful.

You can use it to work out relative performance, but do make sure that you have the avatar in view, it is measuring the direct rendering cost, so be careful about that.

1 hour ago, Echelon Alcott said:

Kudos on the work done. Now people just need to educate themselves and understand how it works, to use it fairly and correctly in its current form.

Indeed so. 

57 minutes ago, Lindal Kidd said:

I don't see it as much different, socially, from Avatar Health > Show Avatar Complexity. That one shows everyone's Complexity figure above their heads, rather than in an ordered list, but it still "hangs a lantern" on high complexity avatars.

The question is, are these numbers more meaningful than the Complexity number? And, how many arguments are we going to have about which figure "really" equates to resource use?

As noted above, you can test for yourself, but render time is quite literally "how long did it cost my CPU to draw this bundle of meshes". ARC is uhm someone waving a finger in the air and estimating based on metrics that are 15 years out of date. 

 

[Beq Janus]

13 minutes ago, Coffee Pancake said:

Your own complexity number is generated by your own viewer. Some viewers fudge the number submitted to SL, for reasons

I think only one does that. and it is not FS

 

14 minutes ago, Coffee Pancake said:

People didn't end up doing anything socially with ARC after it was quickly found to be hot potato garbage, there was an early attempt to police it and eject people from places .. which lasted just as long as it took moderators to change outfits.

Not entirely true, people get jelly dolled and ARC shamed for incorrect reasons, making them make poor decisions on what they should and should not wear/buy. We saw this immediately when the LL perf floater appeared and people were swapping out prim hair for mesh hair cos the complexity was shown to them. A decision that in many cases would have made their avatar slower to render. 

20 minutes ago, Coffee Pancake said:

The µs number, that's actual process time, that's a real number (to 2 decimal places !!)

... which has to be taken in the wider context of what the viewer is doing at the moment, the overall frame time, general graphics card load, capabilities of the computer in question and so on ... the order of the avatars might not be the same for everyone all the time, and could end up very different on different viewers as we all try to reign the avatar processing time .. but hey .. THAT GUY, If we eject that guy everyone's frame rate goes up!!

 

It's a very good, very technical number, with zero context, but if you're over 9000, well, that's actually you.

Yep, I've said a number of times in a number of places now, these numbers are personal to you, they cannot be compared machine to machine,  that is not their intention. the question is "why is my FPS currently X", this tries to give you the answer in a way that can be deconstructed into actions. 

The render time slider is akin to the max complexity slider, but now instead of making you suffer ugly grey blobs on your screen for their vanity, we try hard to reduce their impact without resorting to grey blobs.

My personal favourite though is using the attachment view to measure how good or bad a new hairstyle/head/body is. Self-awareness is a good thing.

One of the good things that has come out of this process is that my initial research into how damning the drawcall overhead of mesh bodies really is did highlight the issue to the Lindens and in the performance viewer we have seen a concerted effort to reduce this through drawcall batching of rigged mesh. the differences are still there, but the magnitude is dramatically reduced and all of SL will win once those viewer changes hit the street.

 

[Paul Hexem]

11 hours ago, Coffee Pancake said:

It would be as just accurate as blaming brunettes for the lag.

Psh. Everyone knows it was the redheads.

[MirandaBowers]

I always found pieces of jewelry to be my biggest offenders for lag. Holy cow they made my complexity jump high a lot of times. 

[Beq Janus]

17 minutes ago, MirandaBowers said:

I always found pieces of jewelry to be my biggest offenders for lag. Holy cow they made my complexity jump high a lot of times. 

Perfect example. They blew your complexity (ARC) high. But did they actually have a signifcant cost? It depends, if they were prim jewelery then compared to the multipart mesh hair or multipart mesh body they would be nothing. It would be an interesting test case.

Edited March 22, 2022 by Beq Janus

[WereBeast Alpha]

I figure this update is better I was able to find out what was causing me to lag was my scripts on my avatar so I took off a lot of huds I didn't need besides my AO and the one for my mesh body. Mesh bodies have high complexity than what classic avatars do. I wish firestorm had more EEP settings that were open.

[MirandaBowers]

8 minutes ago, Beq Janus said:

Perfect example. They blew your complexity (ARC) high. But did they actually have a signifcant cost? It depends, if they were prim jewelery then compared to the multipart mesh hair or multipart mesh body they would be nothing. It would be an interesting test case.

Ahh I see what you're saying. Yes that makes perfect sense now 

[Rowan Amore]

10 minutes ago, xxVi3perxx said:

I wish firestorm had more EEP settings that were open

What do you mean by open?  Look in your Library folder under Environments.

Edited March 22, 2022 by Rowan Amore