The End Of An Era and what was a wonderful, enlightened virtual world....

[Jase Devin]

2 minutes ago, Paul Hexem said:

Your numbers are all wrong.

I won't say they're biased, but they're way off.

I'm running a 2070 with an old CPU and seeing FPS in the hundreds.

As for low end being the bulk of PCs, we have numbers to show that that's not true. Top ten most popular  GPUs in the first hardware survey I found. Obviously there's no way to know exactly what's out in the world, but this is a great sample size.

Why are my numbers way off, where does this come from. I only see roughly 35%, where the other 65%???

Of this 35% about 8% falls under higher end, the rest Intermediate.

[Jase Devin]

3 minutes ago, BriannaLovey said:

My point is that they already do. If you look in the viewer source code, you will see that all hardware info is collected by the viewer and transmitted to LL. They just have no incentive to make the data public. The reasons they gather such data are very different from the reason that the average user would want to see the data too.

Kewl good to know

[Codex Alpha]

4 minutes ago, Jase Devin said:

Why are my numbers way off, where does this come from. I only see roughly 35%, where the other 65%???

Of this 35% about 8% falls under higher end, the rest Intermediate.

We could use Steam Statistics Hardware page if we want a good look.

https://store.steampowered.com/hwsurvey

[Coffee Pancake]

7 minutes ago, BriannaLovey said:

Wasn't talking to you, don't respond to me again.

[Janet Voxel]

1 minute ago, Codex Alpha said:

We could use Steam Statistics Hardware page if we want a good look.

https://store.steampowered.com/hwsurvey

I feel like that’s where he got his numbers from

[Paul Hexem]

3 minutes ago, Scylla Rhiadra said:

Well, great. You go! I'm sure that'll be relief to all of those whose FPS is now running at 4 in their favourite club.

None of us know what the numbers are.

We don't need to know: the evidence is literally right before us that a substantial number of people are struggling. You can't make that fact go away by dropping a hardware survey that has next to nothing to do with SL's user base.

All we have evidence of is that people are whiners. Just because you know someone that's struggling doesn't mean a substantial number are. I know plenty of people that aren't.

You said so yourself- we don't have the numbers. We have anecdotes and ridiculous forum posts, nothing else.

1 minute ago, Jase Devin said:

Why are my numbers way off, where does this come from. I only see roughly 35%, where the other 65%???

Of this 35% about 8% falls under higher end, the rest Intermediate.

That's the top ten from Steam- I could keep going and get more GTX and RTX cards, but it would have been redundant. Your numbers are off for two reasons- both that most surveys out there show 10xx to 20xx being the most common on the Internet, and because a 2060 can run PBR great unless the user is doing something wrong (or at least, not doing something right).

Your numbers are a wild guess based on what you feel like. The only sources we have are third party educated gusses, which is what I posted.

[Coffee Pancake]

11 minutes ago, Paul Hexem said:

Having that data would be better than even educated guesses like the one I just posted, though.

It's honestly less helpful than you would think and the privacy issues are substantive enough that it's not something any of the third party projects will want to touch with a 12 foot pole. Just the very thought "FIRESTORM SPYS ON YOUR COMPUTER" .. wow .. hell no. None of that.

SL is a complicated system, we can have two users on almost identical systems and one is performing fine, the other is miserable.

Raw stats like frame rates also don't paint a contextual picture and everything we do in SL is contextual.

[Extrude Ragu]

I've legit had days where I've logged in, had the viewer pegged at 2FPS, relogged with the exact same viewer and for some reason it's now 90FPS drawing exactly the same thing lol

[Paul Hexem]

2 minutes ago, Coffee Pancake said:

SL is a complicated system, we can have two users on almost identical systems and one is performing fine, the other is miserable.

That is true.

Still, it'd be nice to know how many people are truly suffering from the PBR update without having to rely on wild guesses, biased feelings, or third party surveys that barely apply to SL.

[Scylla Rhiadra]

1 minute ago, Paul Hexem said:

All we have evidence of is that people are whiners. Just because you know someone that's struggling doesn't mean a substantial number are. I know plenty of people that aren't.

You said so yourself- we don't have the numbers. We have anecdotes and ridiculous forum posts, nothing else.

Oh for god's sake, Paul. This verges on literal gaslighting.

Join one of the FS support groups. Join an SL Discord channel or two. Go to Reddit or Twitter or Bluesky.

There is LOTS of evidence of people who aren't "whiners" but are having problems even here.

If you're going to simply shrug it off and pretend everything is just swell, I'm not going to bother with you anymore. We need solutions that will reach a large number of widely dispersed people, not the pretense that the problem doesn't really exist.

[Scylla Rhiadra]

5 minutes ago, Coffee Pancake said:

SL is a complicated system, we can have two users on almost identical systems and one is performing fine, the other is miserable.

Raw stats like frame rates also don't paint a contextual picture and everything we do in SL is contextual.

This.

[Arielle Popstar]

15 minutes ago, Paul Hexem said:

I'm running a 2070 with an old CPU and seeing FPS in the hundreds.

I run a 1070 and what I see on the PBR viewers is sub 100 FPS even in my low prim sky box. I highly doubt that you're getting FPS in the hundreds unless you have everything turned off while looking at the pretty blue sky.

[Janet Voxel]

I feel like the elephant in the room is this build of firestorm is bad.

[Coffee Pancake]

Just now, Paul Hexem said:

That is true.

Still, it'd be nice to know how many people are truly suffering from the PBR update without having to rely on wild guesses, biased feelings, or third party surveys that barely apply to SL.

Easy. Everyone.

Everyone is suffering, but frame rates aren't the singular reason. Neither is base hardware a guaranteed qualifier of a predictable experience.

I can't say "everyone with a 1060 or higher" will be fine, even though for that singular data point they should be.

There are lots of gotchas that were still finding.

One I found the other day - putting a large cache on spinning rust will cause massive stutters and frame rate drops.

[Extrude Ragu]

 

5 minutes ago, Arielle Popstar said:

I run a 1070 and what I see on the PBR viewers is sub 100 FPS even in my low prim sky box. I highly doubt that you're getting FPS in the hundreds unless you have everything turned off while looking at the pretty blue sky.

Something to keep in mind is that a 1070's performance can vary wildly based on all kinds of factors

There are questions like - Is it in a laptop or desktop chasis? Is it being supplied enough power? Does the user have Power Saving mode on their computer? Is it in a smokers house/house with pets.. Cold climate? warm climate? Drivers up to date? Is the owner cleaning off the dust? Was it one of the cheaper 1070's that have worse cooling solutions the list goes on.

[Scylla Rhiadra]

12 minutes ago, Janet Voxel said:

I feel like the elephant in the room is this build of firestorm is bad.

Firestorm has always had a lower FPS than the official viewer, though. I presume that's because it has so many add-on features.

On Friday, as an experiment, while I was at a club (fairly heavily landscaped and meshed up, and also very new) with somewhere between 15 and 20 other people, I did a quick test of four different viewers. I have a newish laptop (8 months or so old); it's pretty fast and generally handled SL pre-PBR pretty well: I could get 80 FPS on my platform, and probably 40 to 50 at ground level.

Using the last FS release pre-PBR, I was getting about 24 FPS in the club. Using the LL PBR viewer, it dropped to about 15. Not great, but actually manageable. Using the new FS PBR release . . . I was getting 4 FPS.

Edited July 15 by Scylla Rhiadra
Typo

[Nalates Urriah]

3 minutes ago, Janet Voxel said:

I feel like the elephant in the room is this build of firestorm is bad.

You ( @Janet Voxel ) and many others think so too. BUT... since my 2016 (i5 GTX1060) hardware runs FS v7 well while it sucks on other computers, even newer, the problem is most likely in the user's settings or their hardware.

For those that want to see if Windows tweaks can improve performance see my blog: http://blog.nalates.net/2024/07/06/firestorm-viewer-7-1-9-problems/

once or twice a month I see my Windows 10 slow down. It has a massive effect on Firestorm. I take care of updates and damaged files and all returns to normal.

[Jase Devin]

17 minutes ago, Paul Hexem said:

All we have evidence of is that people are whiners. Just because you know someone that's struggling doesn't mean a substantial number are. I know plenty of people that aren't.

You said so yourself- we don't have the numbers. We have anecdotes and ridiculous forum posts, nothing else.

That's the top ten from Steam- I could keep going and get more GTX and RTX cards, but it would have been redundant. Your numbers are off for two reasons- both that most surveys out there show 10xx to 20xx being the most common on the Internet, and because a 2060 can run PBR great unless the user is doing something wrong (or at least, not doing something right).

Your numbers are a wild guess based on what you feel like. The only sources we have are third party educated gusses, which is what I posted.

Your evidence supports MY claim not yours.

"Of this 35% about 8% falls under higher end, the rest Intermediate."

[Ayashe Ninetails]

16 minutes ago, Arielle Popstar said:

I run a 1070 and what I see on the PBR viewers is sub 100 FPS even in my low prim sky box. I highly doubt that you're getting FPS in the hundreds unless you have everything turned off while looking at the pretty blue sky.

It's possible he is. A 2070 is still a very decent card. 

With mirrors turned off, a sane draw distance, no shadows, and high instead of ultra, I'd believe it.

Once you cross over into Ultra + bells and whistles + a sim full of avatars, then noooope. Mine couldn't handle more than 40-45 fps like that (I'm on a 3070). Edit: Actually, I'd likely get lower fps on Ultra. I tested it on High.

Edited July 15 by Ayashe Ninetails
Added stuffs

[BriannaLovey]

37 minutes ago, Codex Alpha said:

We could use Steam Statistics Hardware page if we want a good look.

https://store.steampowered.com/hwsurvey

This is like measuring the rate of obesity in Uganda and saying that the same obesity rate applies to the USA

Clarification: What I mean is that you can't draw conclusions about one population by using data from another population.

Edited July 15 by BriannaLovey

[Arielle Popstar]

1 minute ago, Extrude Ragu said:

 

Something to keep in mind is that a 1070's performance can vary wildly based on all kinds of factors

There are questions like - Is it in a laptop or desktop chasis? Is it being supplied enough power? Does the user have Power Saving mode on their computer? Is it in a smokers house/house with pets.. Cold climate? warm climate? Drivers up to date? Is the owner cleaning off the dust? Was it one of the cheaper 1070's that have worse cooling solutions the list goes on.

I was questioning Paul's experience with a 2070 and not my own performance. My posting is not so much for myself rather than my partner's recent acquisition of a laptop with a 1060 in it, which at the time was quite adequate. With the PBR up/downgrade however that Is borderline. Another new laptop is not in the cards for the next year or two and with the other intended updates that are being threatened /promised, continuation in second life for the time being, might not be an option. 

It is easy enough for us to continue on in opensim as at least there we are not forced to change and can easily continue running viewers without PBR. Might be a good thing anyway because at least it gives the SL platform sometime to show actual benefits of the new viewers which we don't currently see.

[Nalates Urriah]

4 minutes ago, Scylla Rhiadra said:

Firestorm has always had a lower FPS than the official viewer, though. I presume that's because it has so many add-on features.

On Friday, as an experiment, while I was at a club (fairly heavily landscaped and meshed up, and also very new) with somewhere between 15 and 20 other people, I did a quick test of four different viewers. I have a newish laptop (8 months or so old); it's pretty fast and generally handled SL pre-PBR pretty well: I could get 80 FPS on my platform, and probably 40 to 50 at ground level.

Using the last FS release pre-PBR, I was getting about 24 FPS in the club. Using the LL PBR viewer, it dropped to about 15. Not great, but actually manageable. Using the new FS PBR release . . . I was getting 4 FPS.

I posted my performance stats for my desktop with v7 compared to v6 here. I lost performance in a couple of places. But I gained in another. My laptop is WAY slower. The laptop uses Intel's Iris HD Graphics. It is designed for streaming more than for virtual world rendering. So I am not surprised.

Having done comparisons for years I know that Firestorm has not always lagged behind the Lab's viewer. They used to trade places now and then. The last few years it does seem to be the case the Lab's is faster.

[Paul Hexem]

8 minutes ago, Ayashe Ninetails said:

It's possible he is. A 2070 is still a very decent card. 

With mirrors turned off, a sane draw distance, no shadows, and high instead of ultra, I'd believe it.

Once you cross over into Ultra + bells and whistles + a sim full of avatars, then noooope. Mine couldn't handle more than 40-45 fps like that (I'm on a 3070). Edit: Actually, I'd likely get lower fps on Ultra. I tested it on High.

Shadows and regular reflections on, SSR and mirrors off, 128 DD, two other avatars, mainland, I can hit 100 easy based on direction.

14 minutes ago, Jase Devin said:

Your evidence supports MY claim not yours.

"Of this 35% about 8% falls under higher end, the rest Intermediate."

"Low end" is integrated, which isn't even 15% of the entire list.

Your claim is, as you admitted yourself, a wild guess based on your own feelings.

[Janet Voxel]

13 minutes ago, Nalates Urriah said:

You ( @Janet Voxel ) and many others think so too. 

I haven’t heard anyone say it. That’s why I did. I’m not going to get into a pissing contest about system specs. I have run the LL viewer, FS and Alchemy. I already knew to turn off mirrors and SSR. In my experience FS runs the worst out of all three viewers. So it’s not about settings. I’m also aware there is some kind of issue with this build with textures. FS still runs the worst.

Since we are putting it all on the table and everybody gets to be heard, all things considered. Why not this thing?

[Paul Hexem]

6 minutes ago, BriannaLovey said:

This is like measuring the rate of obesity in Uganda and saying that the same obesity rate applies to the USA

Clarification: What I mean is that you can't draw conclusions about one population by using data from another population.

Yeah, but unless the TPV devs get around to sharing the data, third party is the only remotely reliable source we have. Everything else is literally just feelings.