NiranV Dean

She made me look through the code. But an interesting tidbit is that the Viewer internally sets the max bandwidth 50% higher than what you set, it also dynamically lowers/raises it depending on whether you are dropping too much packets or not, up/down to 20% more/less it seems. @Alyona Su I also repeated my bandwidth tests and guess what, she's right. Bandwidth ONLY works for UDP (quite an oversight by LL). This means that you should disregard what she said about it since it is no longer relevant, no matter your bandwidth setting, the Viewer will always attempt to download all textures as fast as possible (possibly up to a maximum of 3mb/s if i remember correct). This obseletes what i said before since the Viewer already tries to max its download speed to keep you in download/decoding mode as little as possible. Also, just as little sidenote, i've set my bandwidth down to 100 and 1 to see whether it makes any difference compared to 3000. It doesn't. Nothing changed, the option is 99% useless and for the 1% is could possibly still matter it clearly makes no difference and whatever you set it to is enough (even 1). My point still stands though, you should stop telling people to set their bandwidth down if they are having performance issues, you were only prolonging them unnecessarily, the only ever reason you would throttle the bandwidth is if you are sharing your bandwidth with someone and that someone's share is critically important not to use. And no! Don't give me that "but limited bandwidth provider thingy". Look, the Viewer will download everything, no matter your setting, with a lower bandwidth setting it only took longer, it still downloaded them regardless so ultimately you were only prolonging what was going to happen anyway, it wasting your bandwidth. I'm pretty sure that my technical explanation why you SHOULDNT set your bandwidth down is much better information than what you provided. BTW here goes another round: You won't be able to determine whether it is a system or viewer issue in this case by switching viewers, all viewers drop in performance while textures are downloading and decoding, that is a totally normal thing and no Viewer is different there unless they were to multithread this which no one did so far. All you would be doing is falsely leading them to believe that its a system issue when more than one viewer would show this issue, which will most likely happen given that i've seen people with high end Intel CPU's drop in performance quite hard (this guy being no exception). SO, instead on leading them on false tracks just tell em the truth: This is normal behavior and you can't do anything about it, that's SL.

Downloading and decoding the textures is one of the (if not THE) heaviest task in the Viewer, it is completely normal that your framerate drops while the Viewer is doing that stuff. There is nothing you can do about it and setting bandwidth lower will only increase the time it takes downloading textures and items and thus increasing the duration the Viewer will stay in this low-performance area. You are better off maxing the bandwidth to churn through it as fast as possible.

Anti Mac spew? wth? I didn't say a single *****ing thing against mac.

ah i see, reading it again i just now noticed the "two 8GB"... why don't you just write 16GB... This however does not change my answer, infact it would only bolster what i said, it's going to get worse. BUT It also adds another bit, if you are having issues with 16GB ... are you sure you're running the 64bit version of the Viewer? is there even one for Mac? I don't know how its on Mac but "out of memory" on around 2-3gb sounds like you're running a 32bit version. On windows there is the LAA tag to adress up to 3.5... but even then the Viewer just starts breaking at around 3gb. And if you want to buy a Mac with 64GB RAM you'll need more than one bank. Just sayin.

Excuse me what the *****? I was trying to help and diagnose the problem and regardless of whether you have 4GB or 8GB, my statement holds true. If you have to scrape off as much memory as you possibly can and everything goes boom because one or two apps are using a GB in total of your memory you should consider upgrading, 8GB is little RAM and unless your OS is quite clean and you run absolutely no apps whatsoever you'll have a hard time keeping the Viewer alive, especially now with 64bit. SL is capable of using all your memory now and it will do so if you encounter a good chunk of content. This is a generic issue not just on Mac and neither is it exclusive to Firestorm. If you get your Viewer to 2GB memory usage you will probably not have any issues getting it higher if theres a bit more stuff around and you don't counteract with decreasing the render distance or max avatar count (and even those will only delay the inevitable). I'm not saying anything negative per se, i'm just telling you the hard truth. 8GB is simply not enough anymore, it is more than 4GB (which is absolutely crititical and i wonder how people can even run the Viewer) but as your forum posts show is simply not enough, you clearly have problems keeping your Viewer alive and you semi-actively have to make sure nothing else takes away any memory, that's a big giveaway for too little memory and in the long run this problem will only get worse, creators don't sleep, SL doesn't sleep, both will further develop over time and both will consume more of your already very limited memory and you'll find yourself in this situation again without anything you can do.

I'd assume you only have 4GB RAM, seeing that the Viewer has never before thrown "out-of-memory" warnings is quite lucky and i'd deem that almost impossible, you must have been living under a rock or at least a place where there is never anyone around because those 2gb memory usage you're showing is hit quite easily. Sorry to say it but if 2GB memory usage is a problem for your system you should consider upgrading because this is only going to get worse and its really surprising you haven't had these earlier.

This is hardly a show, i'm dead *****ing serious. I have long given up on EEP. EEP was dead the moment Oz made it clear that they are gonna pull through with this nonsensical marketable item bull***** and empower money driven people to rip off their userbase even more and make windlight unnecessarily more complex and hard to manage and just do what we do.

Hold your horses right there. And the "infinitely better" was solely and totally targeted at the downright broken rendering. EEP is still a massive cluster*****. What i was actually trying to say is "its on the right way" but as always still far from even release candidate worthy. I know i've given EEP a lot of contra in the past and i still do, i'm just trying to say something positive where its due and the progress is better than expected, i truly and honestly didn't expect them to get the overall lighting back so close to the original. I'm sorry. I... was in a good mood. ----- But if all you see in me is pure negativity, sure i can give you the Niran-typical version of above post. The progress is staggeringly slow and the extreme amount of remaining rendering bugs is a shame, its an insult to any person who calls themself a developer that you call this mess a release candidate, even worse that you dare to even think of this being "on a good way" to release, this release is still far from release and the UI is still attrocious and all our concerns about the entire permission system still haven't been adressed, they haven't even been looked into, it would be a lie to say that you are listening to us, all you are doing is damage control on a hopeless project. I have given you examples how to improve the UI, it doesn't take any genius to figure out that the entire permission system is crap and disallowing us to modify any preset at any time and saving it as our own, thus handling mere stupid presets as full blown marketable items is spitting on us machinima people who you are doing this for in the first place.

Gave the latest 6.3.0 another spin. I'm positively surprised, first look looks infinitely better. Lets go through my buglist. (strikethrough is fixed, bold is new) Overall i'd call this some good progress at least the first thing people see is almost the same to what we were used to now. There are still quite a lot of issues to work out tho, all the details and a bunch of fundamental flaws and showstoppers are still in there.

You cant. This is a whole new lighting plus a whole new way all the windlight values are translated, you can't just "opt out". This is a one-way trip.

I don't see how Nirans Viewer was "easier" to use when Nirans Viewer was a lot more technical and offered a lot less help all around, aside from overall offering basically the same UI with less consistent look and much more places "untouched" e.g still Linden UI. I spent a lot of time polishing the usability without dumbing everything down, tooltips everywhere, explanations, a discord group for help, many extra UI features like default buttons, revert to last value, grouping of options, warnings when you are about to do something stupid... the list goes on. Overall BD is exactly the same what NV was just more polished, less of a hastily thrown together pile of stuff.

I highly doubt that, those TPV's probably updated as well.

There was a time where i searched all forums and google for my name multiple times per day. Those times are long gone, i just occasionally search my name and my Viewer's name nowadays and mostly just when im really bored so unless someone mentions me: @NiranV Dean its unlikely that i will immediately respond.

Much like i am doing now? Aha. Are you implying my Viewer is only used because it offers features that are still in development, aka "Bleeding Edge"? That "Bleeding Edge" idea died with Nirans Viewer. What i'm doing now is fundamentally different to what i was doing back then with Nirans Viewer which had completely different goals compared to Black Dragon now. Nirans Viewer was much more focused on continueing the spirit of Kirstens Viewer and sprinkling in some own modifications while Black Dragon is a whole different approach of the same spirit with much more focus on "self-feature-development". Aside from the extremely far and few scattered features from other Viewers, namely RLVa, Exodus Post Processing and Tofu's shader improvements almost everything deviating from the LL default has been done by me, thats the solar opposite of what Nirans Viewer was a collection of features from all Viewers (except Firestorm). Nirans Viewer was a learning experience, it was important for the development of my own coding skills which at the time i started continueing Kirstens were practically 0, i had never seen code back then, whereas now im throwing the "first ever new inworld content creation tool around since a long time in the history of SL" (namely the Poser, btw that's basically what Oz said), that is an extreme jump. Maybe you (or more like Kitty in this context) should feel a bit more honored that i took your "do-it-yourself" approach with things and stopped taking stuff from everywhere.

It wouldnt stop redlining your CPU, just using your GPU more with the same redlining CPU. Since theres always an infinite amount of new extra stuff being shoveled ontop of the work pile for your CPU, being able to work it down faster is not going to help reduce its load, just go through the incoming work quicker, its like a carrot on a stick... except with Vulkan/DX12 its a carrot on a stick with nitro.

Pretty sure they didn't. There isn't much to load off onto the GPU to begin with. In Firestorm there's only a single option i know for sure can instantly 100% your GPU and thats Depth of Field and only if you blur the screen like crazy and have it set to max resolution. In Firestorm you might have overused DoF or Shadow resolution causing the GPU usage to spike so hard up that the CPU starts idling because the GPU is bottlenecking rather then the reverse. Vulkan and DirectX12 wouldn't change a thing about your CPU usage, they would mainly just allow more drawcall throughput which would allow the GPU to work down more stuff, you would just get more rendering done in the same CPU usage (since its the limiting factor in SL). BD has loaded nothing off to the GPU either, your increased CPU usage is completely unexplained but i'd guess its simply because its an Intel. My AMD FX idles most of the time.

It does neither actually unless you have a potato CPU, running potato graphics or found a way i haven't yet to utilize the CPU more than just ~20%, in the latter case i'd like to know how you did that, my CPU would probably be happy to do more than just idling. The most the Viewer has ever used even in full load is the rough equivalent of 2 cores... But if you want more GPU use, you can always just go to a busy place, turn on SSR and Volumetric Lighting, set both's resolution to max, i'm sure they will load your GPU pretty damn well.

It's not like i haven't tried 4 times already. First try was lots of modifications people really didn't like at all and the other three attempts were tries to port back the old V2 style before it was botched with CHUI.

I wouldn't wait for that, that could very well be many years in the future, maybe i don't even get to do it at all. Besides, if anything i would probably try to reimplement the old V2 chat.

Keeping it? Nononono. I just haven't gotten to show it "the chopping block", i have been planning evil schemes to get rid of it since day one. It's tabbed IM after all and i absolutely despise tabbed IM.

Which in intel terms is more like 2Ghz since as we all know Intel has 40% more IPC alone not to mention even the worst dual core Intels seemingly get double my framerates... Let's not forget everyone keeps bashing my FX because "FX's are trash". Point is the FX while having almost 4Ghz is still just as fast as the previous Phenom II X4 i had with 3.4Ghz and still can't even see land against any Intel... not to mention Hyperthreading...

I'm getting 30-80 FPS with shadows on with 8 year old hardware (AMD FX 6200) and a GTX 1060 (i got a 670 before my 670 died but the framerate was the same, GPU upgrades have basically zero impact on your average framerate unless you use GPU heavy features) https://gfycat.com/kindbaggyblackfly I usually never drop below 30 FPS unless the place i go to is an absolute dumpster fire... or i un-jellydoll everyone and have my CPU commit senpuku. But as others said, it highly depends on the place you visit, many objects in SL are simply not optimized (at all), even ripped game assets would be more optimized from the get go. Your settings are usually only a small light unless you touch the three major killers -> Draw Distance / Visible Avatars & Jellydolls / Shadow Resolution

"You never see pictures of the UI because the UI is a horrible mess and completely unusable" ~ that accasional user every now and then. Jokes aside, yes it does use the mostly unaltered CHUI, the only major change (aside from layout tidying up) done to it is the IM tabs being on the right rather than the left. This was done to bring the conversation floater a tad bit closer to the old Viewer 2 chat system which used chiclets (which now don't exist anymore sadly) and autohiding IM windows. The change was done with a seperated local chat in mind, whereas local chat is collapsed to a single line chatbar (like Viewer 2 had) and resting in the lower left corner right below local chat messages while the CHUI conversation window is in the right lower corner (once again like the old Viewer 2) since IM messages appear at the top right and go from top to bottom, this is due to technical reasons, since the CHUI window is no longer "docked" it isn't part of the layout IM toast chain anymore and wont "push" IMs to not occlude them. Here is an example of what is considered the defacto "recommended" default layout. Note that i made the local chatbar not hide on focus loss for this snapshot otherwise it would hide automatically every time i send chat or click outside the window (once again something from Viewer 2 and also Viewer 1), both can be configured seperately.

Well yes thank you Whirly. I had assumed that Firestorm wasn't getting this issue due to being not on the latest version. Apparently my older viewer versions are not having this issue either, can't tell how old since i don't have Windows 10 and i'm not getting this issue (otherwise it would have been fixed long ago already) but i'd guess that going back just a few versions, possibly at the start of animesh (6.0.x) would solve it. People who are reporting this issue are reporting that they have been starting to see this only after they updated the Viewer... but then again its users, who knows if they haven't updated their windows as well. Point is: I am aware of the issue since the very first day and the fix is already planned to be merged, when i'm done with this other stuff.

No, Whirly did that. It was Whirly who got me to test it in the first place. 1) I didn't tell you to not use any color, i asked you to stop using a color that makes lots of people basically unable to read your text unless they select it or switch the forum theme. Besides, setting black as text color is pointless anyway since you are most likely using the light theme which uses black text anyway, setting it makes no sense and only causes issues aside from costing you additional time to write your text. 2) You think that because you are not capable of understanding the issues EEP is having. From our conversation inworld i noticed that you showed no intention to understand the fact that EEP has a lot of rendering issues still, instead you are sweeping them under the rug and say that it is the users simply not knowing how to handle EEP yet. This isn't an issue of not knowing how to handle EEP, EEP's setting differences can easily be fixed when all the rendering issues are fixed, sure they will never look exactly again like they did before, they simply cant but that's hardly the point here. 3) No i haven't. Why should i file a Jira for a hidden side branch back in Viewer 1.1x times when the used code was clearly not meant for public consumption. It wasn't until ~1.2x that it came into the official viewer and until 2.0 that it was acknowledged as an official setting in preferences... and not until much later in Viewer 2's development that it was announced as official complete feature. Unlike EEP it just had instability issues which for a whole new rendering was extremely good, nothing looked broken, flickered, vanished or did other weird stuff. Everything was working, it was just crashing from time to time, something that took many years to get out of the Viewer even without Deferred. 4) Did you, really, expect an unsupported day-cycle length to be properly displayed in the Viewer? This is not a server thing, the server gives you the daycycle preset which is essentially just some basic settings such as length and position as well as the list of used presets for the daycycle, if your Viewer caps at 4 hours, it will do so regardless of what the server says because the Viewer simply doesn't support it, of course it will look broken. 5) I never said i don't like EEP. I'm simply not hyped for it anymore, not since Atmospheric Scattering wasnt to be included due to performance issues. I'm just extremely harsh about EEP because it deserves to. It has a long history that i complain about things but in all those years since i started actually caring about Second Life i've haven't seen anything like EEP, i complain about a lot of things, things that annoy me, things that i think LL does wrong, i do say nice things too... sometimes... if they are deserved of course. EEP so far hasn't earned anything but my criticism simply because there is nothing i see that deserves any good words. The UI is atrociously bad, it is huge, wastes half my screen even when using the day-cycle editor which includes all windows in one (and i've shown just for fun that it could be much smaller) and the XML files aren't better in any way, EEP rendering itself has more bugs than i can count with both hands and my feet and those aren't simple bugs either, they show a fundamentally broken rendering everywhere. What takes the cake however is the fact that LL was considering to release this in such a horrible state because apparently everyone here was just sugarcoating EEP like it was Jesus's second coming, when i scroll back the pages and look through reported issues i see lots of "horray this is so awesome, wow look at what i made" and a little "hey this looks a bit weird" sprinkled inbetween, no one seemed to have raised concern over its actual horrible broken state and that started to concern a few other people who actually care, namely Whirly who contacted me and asked me to look into EEP. When Whirly is so concerned that she asks ME to look into it and give it my ranting then YOU should be all the more concerned. I don't hate EEP and i don't hate everyone who is involved in it, i'm just super concerned that EEP is going to be released as the *****show it is right now and that would not only mean that your windlights would break, no... rendering would be totally broken for everyone and it would be live on a large amount of users, this would create a *****storm of epic proportions... that is if people even notice. Seeing how little people here saw the issues before i pointed them out worries me even more, how many more people would see these issues if it went live? It would be a catastrophy if this broken rendering would become the new "default". So IF ANYTHING, i'm extremely worried about EEP. 6) Great you had to bring it up. Look, my Viewer is usually one of the first to adapt recent LL Viewer changes (aside from Cool VL), just because i don't have a fix yet that was just recently released a few days ago doesn't mean that i'm standing in a glass house throwing stones. It's the total opposite i'd say, the fix is already planned to be included in the next update which just happened to take a lot longer due to ... technical problems because unlike EEP i wanted to get my ***** right, which is why i sat the last 9 days alone working on these goddamn profiles day and night and i'm still not fully done, they are working, they look nice and should provide almost if not complete functionality, they just need some touch-ups and maybe a fix here and there which i missed, which will come later. I haven't exactly been sleeping the past few weeks, i've already said on multiple occasions that the fix (as soon as it was released as RC) will be included in my next update which will be soon. But and that's something you should never ever do again, you seriously compare my Viewer to that of Firestorm and say that even Firestorm has it fixed? Let me tell you something: Firestorm whenever a new release comes out is outdated, every, single, time, their current version is still on 6.0.2 which didn't have this issue YET, the issue only happens in later versions in conjunction with the latest Windows 10 Update 1903. The windows update 1903 came after said 6.0.2 update, so at the point this issue surfaced everyone who wasn't lucky enough to be on older versions was screwed, lucky for Firestorm they were outdated, their version and all older ones are not subject to this bug. Bad for me because unlike Firestorm my Viewer is more recent so it was hit by said bug and when said bug surfaced i was in the middle of a huge rework which i couldnt just drop, so i had to power through this the past 2 weeks to get it done in a decent fashion and FINALLY give my users the update they need to have this bug fixed. So don't tell me i'm sitting in a glass house throwing stones, i have been hard at work making sure i don't do the same mistakes i'm constantly ranting about here but it's hilarious to see you bringing up this issue in conjunction with my Viewer, you know if i wasn't stuck in this rework hell i would have had a fixed released 2 weeks ago already but theres this one time Niran takes a bit longer than usual and suddenly the stars align and someone jumps out of the corner and points at me how bad i am for not having a release ready with a fix that was just released. Oh boy.