Reporting Viewer Performance: Raising the BS threshold

[Qie Niangao]

In a recent post to her blog, Inara Pey gives results of an impressive array of tests of several different viewers.

The testing procedure, she admits, is not perfectly rigorous, but it's way better and much more comprehensive than other reports.

I'm hoping this gives folks the ammunition to tell others to go get stuffed when they insist  "everyone should use this viewer cuz it's faster on my machine, I can just tell."

I'm also hoping the really dismal showing of most TPVs compared to LL's Viewer 3 will motivate some of those developers to do some real benchmarking.  (The report already has them squealing like stuck pigs, but we'll see if they'll bestir themselves to collect actual data on more than their own homebrew viewers at idiosyncratically optimal settings.)

[Imnotgoing Sideways]

I like the direction V3 is going.  Honestly, I miss the sidebar a bit.  I always liked tabbed IMs and grew to prefer tabbed inventory/group/whatever dialogues.  I am also extremely anti-pie menu.  2/3rds of my computing day is spent in Visual Studio, MS Office, Chrome, and SolidWorks... Each of which drops a contextual dialogue which I can slide right into.  Suddenly faced with a "pie" menu, I'm looking for things in a circle.  But, what I'm looking for isn't there. Click "More", still not there. "More" again, still not there... Until I've gone full circle and finally find it on the second "More" page where I overlooked it the first time. (=_=)

So, no, you can't PAY me to use any throwback UIs. (>_<)

That said, I currently use Niran's (Which happens to have a pie menu option... that I disable).  I can surely attest that the frame rate is near the low end of all clients I've used before.  But, I found that I'm pretty much incapable or crashing it on my system.  In addition, I can turn up every graphic setting and take a high resolution screenshot without immediately going kaput.  I can't dislike that. =^-^=

When asked which client I'm using, I half jokingly respond "Not the one I was using yesterday".  Because I somewhat constantly rotate, try out betas, try out dev builds, and anything else that comes down the pike.  I will never say any one client I use is "better" or "best" in any way.  But, I can say which is my favorite.  At the moment, it's Niran's.  Now that I've seen these FPS comparisons, I'm going to try some others out to see if anyone else has achieved the stability I find in Niran's with the added bonus of some FPS on top. ()

Oh... And it MUST render mesh.  I'm having too much fun toying with the stuff. =^-^=

 

 

[Alicia Sautereau]

As of this morning with the latest release of Exodus, i`ve removed all traces of the other viewers

It has a TON of extra options along with extra visual effect (that need abit of tweaking per system settings)
It`s just brilliant for the higher end pc`s where you can put everything on high(client & gpu control panel) and enable all the visual extra`s

There are still bugs, but it`s beta

[Rene Erlanger]

I thought Ukranian coders were used for development of  Viewer 2....commissioned from it's SF head office! I've no clue who worked on Viewer 3's.

[Qie Niangao]

Yes, the Ukranian outfit (productengine) may still do some viewer development, I'm not sure.  To pin responsibility on somebody for the worst design misfeatures of the original Viewer 2, however, I think we need to look closer to home.  (If we want to engage in conspiracy theories, we could pretend that some board member's idiot nephew is one of those ever-so-trendy-looking consultants.)

FWIW, I spent some quality time with the latest exodusviewer, and was pleased to see that it appears to fare pretty well in comparison to all but the Linden viewer... but my main point remains: Those who make viewer performance claims now have a minimal standard of measurement, and can put up or shut up, as far as I'm concerned.

[Perrie Juran]

---

Qie Niangao wrote:

In a recent

post to her blog

, Inara Pey gives results of an impressive array of tests of several different viewers.

The testing procedure, she admits, is not perfectly rigorous, but it's way better and much more comprehensive than other reports.

I'm hoping this gives folks the ammunition to tell others to go get stuffed when they insist  "everyone should use this viewer cuz it's faster on my machine, I can just tell."

I'm also hoping the really dismal showing of most TPVs compared to LL's Viewer 3 will motivate some of those developers to do some real benchmarking.  (The report already has them squealing like stuck pigs, but we'll see if they'll bestir themselves to collect actual data on more than their own homebrew viewers at idiosyncratically optimal settings.)

---

I AM AN ABSOLUTE SUPPORTER OF "YOUR WORLD, YOUR CHOICE OF VIEWERS"

And I also agree that people should have concrete data, not just vague impressions. 

But it seems to me that you have chosen to ignore what she stated at the end of her blog post:

"So, in conclusion, you’re free to interpret these results as you see fit; how much value they represent is questionable. As always, individual experences may vary wildly from my own (particularly those of you fortunate enough to run a higher-specfication CPU / GPU combination). However, as a finger-in-the-air reference point for my own reviews, the tables may have value, and I may maintain them…

Again, to be clear: I’m not claiming the test is designed to be either empirical or scientific – please do not take it as such."  (emphasis added)

If you want to see screen shots that offer a completely different and opposite experience than hers, check my JIRA here:

https://jira.secondlife.com/browse/VWR-27983?page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel&focusedCommentId=304027

SL Official 3.2.4 ----- 5 FPS !!!!

Firestorm 3,2,1 ----- 10 FPS

Phoenix 1.6 w MESH ----- 16 FPS

FireStorm Beta 2.5.2 ----- 19FPS

Sorry Qui, but all you did was raise the threshold higher.

[Qie Niangao]

Oh, did I not say that the testing procedure, she admits, is not perfectly rigorous?

Why yes, I believe I did.

In any case, nothing would make me happier with this thread than if, indeed, all I did was to "raise the threshold higher" for reporting comprehensive viewer performance metrics. That's precisely the point.  And if one result of that is to get some firm grasp on specifically why some installations seem to generate much different results, that would be a win.  Wouldn't it?

[Perrie Juran]

---

Qie Niangao wrote:

Oh, did I not say that the testing procedure, she admits, is not perfectly rigorous?

Why yes, I believe I did.

In any case, nothing would make me happier with this thread than if, indeed, 

all I did was to "raise the threshold higher" for reporting comprehensive viewer performance metrics. That's precisely the point.  And if one result of that is to get some firm grasp on

specifically why

some installations seem to generate much different results, that would be a win.  Wouldn't it?

---

It was this quote that I am objecting to:

 

---

Qie Niangao wrote:

I'm also hoping the really

dismal

showing of most TPVs compared to LL's Viewer 3 will motivate some of those developers to do some real benchmarking.  (The report already has them

squealing like stuck pigs

, but we'll see if they'll bestir themselves to collect actual data on more than their own homebrew viewers at idiosyncratically optimal settings.)

---

It is your use of these inflammatory terms, "DISMAL" and "SQUEALING PIGS." and your general accusation that the TPV developers don't do any "real benchmarking."

Based on my results I could say that I have a pretty good argument for lack of "real benchmarking" on LL's part than the TPV's  though actually, I prefer the term 'quality assurance.' 

Regardless of which term you use, there are so many different system configurations out there that 'benchmarking' is near impossible to do. 

While you may not like the fact that there are people who for whatever reason do not like the Official Viewer, be it performance or interface or some other vague reason, for you to besmirch people who for the love of SL or of Coding, who don't even get paid for their efforts to give us a 'Better World,"  I consider that very uncool.

You came in here and tried to build a case based on one persons admitted unscientific study.

That just doesn't work.

[Qie Niangao]

Heh.  Well, see the comments to Tateru's blog post on the subject.  "Squealing" is not an ungenerous characterization.

Maybe that TPV dev has real benchmarks that mean something beyond his personal machine, maybe he doesn't.  Your results, and Inara's too, may or may not generalize.  The point remains, however, that a few uncontrolled datapoints as a benchmark "just doesn't work" any longer.  Whether you agree with Inara's results or not, she's raised the bar for measuring and reporting on viewer performance.

Also as another cited in a comment on Tateru's blog, LL does very extensive performance testing on different GPUs.  It's not cost-effective for them to run such comprehensive tests on every release, but it's not as if they never do viewer performance testing.

[Perrie Juran]

---

Qie Niangao wrote:

Heh.  Well, see the comments to Tateru's blog post on the subject.  "Squealing" is not an ungenerous characterization.

Maybe that TPV

dev has real benchmarks that mean something beyond his personal machine, maybe he doesn't.  Your results, and Inara's too, may or may not generalize.  The point remains, however, that a few uncontrolled datapoints as a benchmark "just doesn't work" any longer.  Whether you agree with Inara's results or not, she's raised the bar for measuring and reporting on viewer performance.

Also as another cited in a comment on Tateru's blog, LL does very extensive performance testing on different GPUs.  It's not cost-effective for them to run such comprehensive tests on every release, but it's not as if they never do viewer performance testing.

---

Here now you are talking about ONE TPV's  Developers comments.

In your original OP you stated:

---

Qie Niangao wrote:

I'm also hoping the really dismal showing of

most TPVs

compared to LL's Viewer 3 will motivate some of

those

developers

to do some real benchmarking.  (The report already has

them

squealing like stuck pigs, )

---

using plural nouns.  I only see ONE TPV responding in Tateru's blog.  So hence again, your original OP was and is inflammatory.

And as far as that Developer "squealing like a pig" goes,  that  is your interpretation and having read through his comments I would also say your misrepresentation of what occurred there.  What I see is the TPV logically responding to the issue.

In addition, for Tateru to say

"On the whole, though, while you might not get the same numbers as Inara, you should get results that are consistent relative to each-other as she does, given the same tests."

Is a complete assumption on her part unless she has multiple other test results to cite from.  Quite obviously, my test results show that her statement may not be true.  So if I were that TPV, I'd be responding the same way also.

Tateru stated the tests were unscientific and not reliable but then says you should get about the same results?  Sounds contradictory to me. Where is the research to back up such a statement?

I have never heard a TPV say "my viewer is the best that there is."  What I have heard them say is "here is what my Viewer has to offer," though I wouldn't doubt that some of them think that.  Personally, I think the TPV's have done a better job at customization than anything Linden Lab has done.  But that is just my opinion.  I still tell people "Your World, Your Choice Of Viewers."  Use what works best for you.

And on a final note, just to be clear, I never said that LL didn't test things.

[Sigren Panthar]

You came. You wrote. You failed.

Let me just ask you this...as a apparent and upcoming shill, it has clearly never occured to you that people like myself do not use the dire LL viewer out of spite or an act of rebellion; we don't use it because it isn't good enough.

To be fair to LL, the direction the new viewer 3 is taking is - at long, long last and after years of resident complaints and pleading - going in the right direction.  However, for a company like LL to bring out the disaster that viewer2 was in the first place, and then bury their heads in the sand and pretend all is well was nothing less than a scandal.

I do not use viewer3 as my TPV frankly p***es all over it.  Not because I have any particular loyality to any TPV dev team.

Now, do you grasp this concept?  If the LL viewer is ever as good - and, again, to LL's shame their OWN viewer should be the best in theory bearing in mind their resources, but to their eternal shame it is not - and better than a TPV viewer, I, and many others will embrace it.

Your blog link/ test result means nothing. Absolutely nothing. Why? because people try different viewers and pick the best one for themselves to enjoy SL with, and for the majority of residents it is not viewer 3.   That may cause a lump in your throat, but that is a fact.  Accept it, until LL delivers what it should have long ago.

 

 

[Qie Niangao]

---

Sigren Panthar wrote:

shill

---

These forums, they break me up, they do.

[leliel Mirihi]

---

Sigren Panthar wrote:

You came. You wrote. You failed.

Let me just ask you this...as a apparent and upcoming shill, it has clearly never occured to you that people like myself do not use the dire LL viewer out of spite or an act of rebellion; we don't use it because it isn't good enough.

---

So if I understand you correctly, anyone that posts something positive about the official viewer is shill, right?

---

Now, do you grasp this concept?  If the LL viewer is ever as good - and, again, to LL's shame their OWN viewer should be the best in theory bearing in mind their resources, but to their eternal shame it is not - and better than a TPV viewer, I, and many others will embrace it.

---

I think you're the one that's having trouble grasping something. The blog post was solely about benchmarking viewer performance, it didn't say anything about which viewer had a better UI.

---

Your blog link/ test result means nothing. Absolutely nothing. Why? because people try different viewers and pick the best one for themselves to enjoy SL with, and for the majority of residents it is not viewer 3.   That may cause a lump in your throat, but that is a fact.  Accept it, until LL delivers what it should have long ago.

---

NO SH*T? You don't say? Who would have thought? The blog post you're deriding even mentions at the end that despite their lower performance some people may prefer [some] of the TPVs due to their extra features. 

I could go on for a while longer about how biased you are, but really what's the point? You're just going to label me as a shill and dismiss my post out of hand without even reading it, despite the fact that I never said a single positive thing about the official viewer (or a negative thing about the TPVs).

[wiked Anton]

actually, it is only in her opinion that those are the fps rates. i find firestorm laggy, and jerky on my computer, i only have a lower end, but dual core with a big graphics card. i find phoenix much better for my machine. anyone saying a particular viewer is better than the rest in general, has a biosed opinion and will be ignored by me.......

[Ceka Cianci]

Oo *Sighs*

My post was removed? Seriously?

Do they still have that STUPID forum feature where if someone deletes their post all posts tied to it are deleted as well?

 

[Ann Otoole]

I have not seen any productengine people posting commits on the dev mailing list in a long time. But I wasn't specifically watching for them so could be wrong. However my recollection is the only activity is open source dev contributors lately.

[Melita Magic]

The last thing Qie ever is, is a shill, or any other form of fraud. The person who said that is talking out their pixel backside.

[Rene Erlanger]

Why did the moderators deleted half this thread....along with 3 of my posts.?  One was asking for a suitable viewer for my machine...the other 2 were about Cool Viewer being very good during its trails....and that i was willing to try out LL's V3 Viewer.

Jeeze been away from SL for quite some time....only to come back to heavily moderated Forum! :matte-motes-sour:

[Theresa Tennyson]

My understanding is when a post is removed posts that reply directly it are automatically removed as well, and that the original poster can remove a post which will also remove replies. Some of the posts on this thread were concerning ethnicity and were becoming heated. Your posts probably got caught up in the crossfire.

[Ceka Cianci]

---

Theresa Tennyson wrote:

My understanding is when a post is removed posts that reply directly it are automatically removed as well, and that the original poster can remove a post which will also remove replies. Some of the posts on this thread were concerning ethnicity and were becoming heated. Your posts probably got caught up in the crossfire.

---

Oh i just hate getting caught in the Crossfire..

i just makes me have to do

THIS!!!

.

 

[Ann Otoole]

I see Andrew from Productengine changed something from QA back to review today. I guess they are still around.