Nalates Urriah

Yeah... we blame everything on her... 🤣

Intel's HD Graphics and UHD Graphics will run games. But not well. The system is designed for streaming media. Think movies. A quick way to figure out which HD Graphics is new and better is to look at the CPU. A i3, i5, or i7 comes in generations, 1 to 11. Each generation has a better graphics engine. You can tell which generation by the chip designator, i5-6600k. This shows it is the i5, a mid-performance model often considered as giving the best price-performance ratio for gaming. The 6600 tells more about the chip with the leading 6 being the generation designator, so 6th gen. The 'k' suffix tells us it can be overclocked. With even the best UHD Graphics you'll likely run SL on a Medium or Medium-High setting. Above that and you'll likely pull frame rates into the single digit range.

There are avatars that NEVER change... even clothes. Several of the Lindens have an avatar that does not change or at least not often. Several of the developers/designers that attend the User Group meetings almost never change. But, there are people that go into Coldstone and order vanilla...

Firestorm recently (7/29/2020) released their beta version of the viewer with EEP. Version 6.4.5. If you recently updated to that version then the change is in the EEP rendering. If you use a version back (6.3.9) and do not have that problem then it would be good to let the FS Devs know. But, make sure the same hair-base is rendering differently in both versions.

You can use the free program HWMonitor to watch temperatures in the computer components. MemHistory is another free program that will show you how memory is being used. You'll be able to see if the 8GB is being consumed. The clues to what is happening are in the viewer's log file. See this post to find the log and make some sense of it.

Have you asked someone in the FS Support in-world UG to walk through this with you? Also, this type of question is best asked in the Answers section of this forum.

I've forgotten which viewer you use... if I ever knew. In LL SL Viewer try Me->Movement->Stop Animating Me. In Firestorm try Avatar->Avatar Health->Stop Avatar Animations. When you relog move to another region. Relog again. Does it still happen? Read the error dialog carefully. May be post a screen capture of it here. You can copy paste into the forum. Once a session ends things you have taken off don't have an effect. So, look closely to make sure you have gotten rid of all the parts of whatever that may have got left behind. The quick way to do that is select an avatar from the Linden list of starter avatars. Then put yourself back together an item at a time.

I had no clue what a MINIS is... Give us some numbers. Which viewer are you running? How long does a never visited before scene take to render? Pick a well know place like Tralala Dinner. What type of FPS do you get in a Safe Hub with 20+ people or at an event like Cosmopolitan? What FPS do you get in an empty region at 1500m altitude? I ask because from the specs it looks to be really slow.

There are some threads here about using duel video cards with SL. Plus there are some how to's around the net. I am not sure what you mean by duel GPUs. It is a bit hard to count GPUs these day as Nvidia says, ...GPU has all 2304 shaders enabled, NVIDIA has disabled some shading units on the GeForce RTX 2060 to reach the product's target shader count. It features 1920 shading units, 120 texture mapping units, and 48 ROPs. Also included are 240 tensor cores which help improve the speed of machine learning applications. The card also has 30 raytracing acceleration cores. But I suppose that is a single GPU to most of us. We usually refer to duel GPUs in SL as being two cards. So, two 2060 cards or 1060's or some pair of cards. Intel has a GPU section built into their CPU chips. They say it this way, Intel manufactures the -9600K on its 14nm++ process. In addition to six execution cores (without Hyper-Threading technology), the chip includes an integrated UHD 630 graphics engine, sports unlocked ratio multipliers for easy overclocking, and supports two channels of DDR4-2666 memory. I suppose we could call that a GPU. But it is within the CPU chip and uses system RAM not having dedicated memory like an AMD or Nvidia card does.I think it likely to be near impossible to get that 630 engine and a Nvidia/AMD card to cooperate and act as duel video processors. Even with two Nvidia cards cooperating one does rendering and one does support as the viewer's render process is described as a single thread process. So... what do you mean by duel GPUs and which ones do you have?

Well... the fact is all display devices are pixel based and that requires integer math. You are good at voicing your opinion but no facts. You just criticize me and attempt to be insulting. At least I corrected the mistakes I had in my opinion.

We disagree. But, at best the real answer is more complicated then we are saying. This article from 2003 gives an overview of the thinking then. Notice the section on Floating points. https://www.flipcode.com/archives/3D_Graphics_on_Mobile_Devices-Part_2_Fixed_Point_Math.shtml - Yes they are talking about mobile. But Googling for more information it should become obvious that most games of the era are using fixed point math, which is integer math over floating point (ref 2010+/-). That is what I was playing with in high school when I was hacking around in the Myst Online game and formed my idea games use integers as they are faster. Molly makes the best argument in defense of your position by referring to the OpenGL calls. It is at 4096m +/- altitude that the precision of SL math most definitely and obviously breaks down creating visible problems. Above 3000m people see some problems. Those designers that have not put effort into keeping the vertices of their models limited to near integer values find seams opening in their model. It is a common problem in SL. We can go above 4096 but the precision of models falls apart and we can see problems in most models. This is common in fixed point math models, which is why I thought SL ws using integer math in the render engine. But, that was supposition on my part. I did start looking at the Linden Viewer code to see if I could quickly find an answer. Then I realized from Molly's comment that the working part of the engine is OpenGL. That is another load of code to look through. Fortunately there are discussions of how to best use OpenGL that show up in searches. This is where I find the articles explaining how the engine converts 3D world values (floats) to 2D screen coordinates and how the floating point values passed into OpenGL get converted. Then I wondered about my thinking that integer is faster than floating point. A bit of research on that reveals that in GPU territory 32-bit integers and floats will process at the same speed. With that in mind, a quick look at OpenGL shows a mix of floats and integers in the processing. But the direction is ALWAYS toward integer math to render to the screen. Well... duh! The screen, any screen, is pixels and the location of all pixels is represented by integers. I should have remembered that point and started there. So while the SL side of things can be all floats as they pass through OpenGL to screen for display they have to convert to integers. That will true of the newer Vulkan too. So, while I was thinking SL Viewer code would use integer math for rendering, I was wrong. They don't have to. In fact the Linden code could be said not to do the render but passes it off to OpenGL. They use a a mix of types in the viewer depending on what they are doing or is being represented. OpenGL does the same but with the twist that they eventually have to get to integers to put the result on the screen. So, most... no... ALL render engines use integer math. How much they depend on it or make use of it is debatable. Reread your post. Do you notice that you begin attacking me rather than providing a critique (sources) based on facts and references?

Well sort of a comeback... The BOM change allows FMBs to work more like the Classic body. You can layer clothes on a BOM-FMB just as you did the Classic body you are used to. Since it is system clothes being layered, you can make them just like you remember. However, I think system clothes are now used mostly for skin, tats, underwear, and shirts that need to fit under mesh clothes. It is the mesh bodies, heads, and clothes that are popular. The Classic body is not all that popular nor making a comeback. But... Onions have many layers. The FMB pre-BOM was made like a set of Russian Dolls. One inside another. The inner most copy of the body was for skin. Since prims can only have one texture/diffuse image and cannot be composited with other images the tats layer was a copy of the body just a tiny bit larger than the skin layer. The same with the underwear and clothing layers. Pre-BOM bodies had four times as many polygons and textures as needed by a Post-BOM body. BOM is a Linden effort to reduce the render and data load on the system and it reverts us back to how Classic bodies were dressed. So, I would say it is System Layers that have made a comeback.

Most real-time render engines still work with integer math. They have hundreds of billions of operations to perform. Even small performance gains multiplied by billions becomes significant. Floating point numbers in computers are used for those things which require them. Even our desktops have dedicated floating point processors in the CPU. Video cards are made for other purposes than just graphics display, think bitcoin mining. While numerous biology, physics, and chemistry problems need a floating point base, 3D rendering doesn't. Integer math measured in MIPS is very difficult to compare to Flops. Deciding which way to go in scientific circles usually depends on the size of the number and the precision needed. Will 8 bit numbers get by? 16? 32? 64? 128? The more bits required the slower the computation. So on modern desktops with dedicated video cards it may be possible to get acceptable performance for an app like SL using either. The apps needing larger numbers tend to use CUDA processing in GPUs where there are hundreds of not thousands of GPUs in the card. The actual limitation for SL is the size in bits of the numbers that give just sufficient precision and the best speed. In LSL 32 bit signed integers are used. I would have to dig into the render engine to see what is actually used there. I suspect 32-bit unsigned integers were used. The render math was decided on in 2003-4. Now it is too big a task to make a change. A change would sort of be like changing the foundation of a high-rise building so you could add more stories. So much stuff has to change it is easier to tear the building down and build a new one. Which is why the Lab decided to build Sansar rather than remodel SL for VR. Now it looks like the hardware may get fast enough that SL could possibly run in VR. But a lot has to improve get to a minimum of 90FPS. I should have written 4095 decimal = 1111 1111 1111 binary. Stopping to do a bit of research I find http://wiki.secondlife.com/wiki/LLSD#integer But, it does not talk about the render engine.

You are coming back in the middle of a transition. Your old Classic avatar is very viable and the way we work with tattoos, skin, system clothes, makeup, and all the things you are used to are now the 'new' way. Sort of having come full circle. While you were gone mesh was introduced from that came standard size mesh clothes, Liquid Mesh clothes, and now Fitted Mesh clothes. Since we weren't getting an Avatar 2.0 designers came up with Fitted Mesh Bodies (FMB). But FMB did not use the system layers you are familiar with. Stacking skin, tats, underwear, and clothes as we do with the Classic avatar was not possible with FMB. So, onion skin FMBs were developed and appliers became a thing to put a tat on a tattoo layer, skin on a skin layer, etc. Now we have BOM that allows a FMB to have a single layer, no onion skinning, and use system layers to stack things just as the Classic avatar does. So... we no longer have to use appliers. During this transition there are lots of old FMB onion-skin style stuff still around. Onion skinning is still common with heads. One has to pay attention to which method the designer is using with there products. So it is going to be confusing. But, what you knew is now back in fashion. I made a post to help another returning person understand the new SL jargon in older SL terms.

In the viewer tweak the ambient light to get more contrast. Depending on the viewer you use it is in different places. Firestorm has Phototools which is the easy way to make adjustments. The is D button for default so it is easy to restore settings. Play and experiment. If you aren't into tweaking individual settings, use the Windlight/EEP environment settings to quickly change things.

@Jennifer Boyle The short answer is because of creative freedom. Understanding that gets complex. But the complexity is the reason it going to be practically impossible (impractical) to provide the software you want. Think about it. A mesh head is a series of vertices that describe the head. Each vertex is a X, Y, Z representation of a point in 3D space. Some thing like <124.15, 216.28, 30.22>. Modeling tools save the modeler from dealing with the actual numbers. They visually sculpt the head. So, each head has a different number of vertices with each in a unique location. If the vertices were all in the same place and same number of them then all heads would look the same. The next level of complication is in weighting the vertices. The head needs to have eyes that blink, a mouth that opens, lips that smile, etc. To do that we have to tell the system how each vertex is to move in relation to a bone, say the jaw. The values we call weights are the information that tell the system how closely a vertex follows a bone. In SL there are four values that tell the system how the vertex moves in relation to four possible bones. These values are also handled by modeling software and allow creatives to paint with a brush to weight masses of vertices. They don't really consider the numeric values. The point here is it is very impractical to expect designers to strive for numeric consistency between various heads. Without that consistency the software you want would have to know all the values used in each head to be able to cross translate a shape from one to another. Making that even more impractical is that vertexes in nearly the same place in a head are weighted to different bones. This is why some animations don't work well with some heads and others do. To eliminate all the complications would require removing creative freedom from the design process. That would make all heads look pretty much the same. Not what most of us want. It took me a long time to find a mesh head I could make look similar to my classic head. There are a large number of heads I just could not tweak to the look I wanted because they looked different. They could only be tweaked so far from their basic look. The shape side of the viewer deals with bones. The shape editor has no knowledge of the weight values and vertex positions. When you tweak a shape you are eyeballing what moving the bones does to the mesh without any need of all those math values. The use of those numbers is buried deep in the render engine. Most of us never need to know anything about them.

Lots of opinion and speculation here... The simple fact is all digital computer models use mathematical descriptions of the things within the model. If I alone am working in the model it is I alone that needs to be consistent with my self. Anyone can decide the units in their system are whatever they want. Blender, AutoCad products, Draftsight, and other modeling programs will allow you to designate the software's UNITS as what ever you want. People regularly change those units. Someone wants to represent the galaxy, another the solar system, another a country map, another a house, another is designing an engine or a smart phone... or modeling a virus or protein string. Because modeling software is designed to be fast the mathematical representation is usually integer based. Only when people start adding dimensions does floating point math get involved. If that seems odd read some Wikipedia. It is when we start to work and cooperate with others we have to decide how we will think about the units we use. The Lindens at some point decided a unit in SL would be a meter. But the math is such that at 4000.0000m is the limit (4098 = 1 0000 0000 0000 b) where precision breaks down. It is a matter of how the SL integers are translated to floating point values. A distance of 200.003m is a floating point number. But the render engine sees it as the integer 200,003. This should explain why using s dimension of 5.0014 gets rounded off oddly. So, the base unit in SL while labeled a meter the actual unit is much smaller, like a millimetre. Confused yet? The point is the units in any modeling program are purely arbitrary. SL has a standard because of the demands and limits of 2003 software and computers and a need for people to be able to work together.

I have had mixed experience sailing and flying in Agni. In general it has been better than 2 months ago. I have sailed for hours with no more than the 2 to 7 second delay at crossing. That is with a Bandit 55, ACI 90 + 3+/-MB of personal scripts. I often sail in Tuesday's Topless Cruise. This last Tuesday I was crashing. It was a week without main region restarts. The week before, a restart week, I sailed the course with only one crash. Two weeks ago I flew in Eden-Fruit Islands making 91 crossings without a problem. That was with a Shergood copter, my mesh avatar and 3+/-MB of worn scripts. It appears people are having different experiences. I was curious how the Aditi cloud regions compared. IMPRESSIVE. You can see the video I made playing with a Bandit 380R here: http://blog.nalates.net/2020/08/06/second-life-region-crossings-2020-whats-coming/ From todays UG 8/11... [12:10] Simon Linden: the better region crossing code was rolled today and with tomorrow's update it will be on the full grid. It's NOT moving simulators into the cloud, but was discovered with cloud work ... if you think old region crossings were bad, the ones between our SL servers and the cloud were much worse. Fortunately it clearly pointed out a problem, and the fix turned out to be even better than what we had previously Crossings were on the order 0.2 to 0.5 seconds.

This is an old Grendel's dragon. I thought it too cute to pass up.

You can ask these types of questions in the Answers section of this forum. In the case of your question the answer is hard to find. In personal profiles you can use the sort of generic BBC and forum style link. Use [ the url then a space then the text you want then close with ] [https://community.secondlife.com/forums/topic/459315-in-world-profile-question/ A Great Link] The forum won't convert this text to a link. But it should in your profile. AFAIK this does NOT work in group profiles.

There are free ZHAO II AOs. You can move your animations from the free AO into most other AOs. Your change should be free.

That is awesome. Thanks!

WE ARE! But you do it so well! ...of course in many cases we were seriously disturbed before we met you.

Many of us are attached to our SL appearance. And yet over time we change. I think the more attached we are the slower we change. But most do change. The step by step changes... There are threads around here where these types of images are the thing. It helps to have good images of your avatar. Getting some each year pays off when these memes come up.

Some people are thin skinned. In a world with racism, cancel culture, anarchist, violent protesters, a corporate media that disses anything counter to The Agenda...and they think think this forum is too much to handle... 🙄 they probably should stay away. A Dislike Button would get too much use. SL has lots of jerks, plenty of the insensitive, loads of ignorant, many that are dreaming they are the woke, the intolerant, anti-diverse, and just plain haters who would be all over such a button.