Jump to content

Ansariel Hiller

Advisor
  • Posts

    3,276
  • Joined

  • Last visited

Everything posted by Ansariel Hiller

  1. DRTVWR-563 is going to (try to) address this and will use all available VRAM, using methods of the operating system to determine (or guesstimate if you are a poor Mac user) the current VRAM usage. There will be no texture memory slider anymore - which also means Linux users should prepare to get bent over hard unless some Linux fanboy implements those VRAM detection method on Linux... 😁
  2. This can usually be caused by two things: Microsoft-provided GPU driver is used Viewer is using the integrated Intel GPU instead of the dedicated nvidia GPU Make sure the driver from nvidia (or the nvidia driver provided from Dell) is installed. And also make sure the viewer is using the nvidia GPU (usually assigning a high performance mode for that application).
  3. ...prepares next "slower than other viewers due to million options because there's always somebody who can't live with the slightest changes and demands options plus megabug" release of Firestorm...
  4. First: "U" is not a word, but a vowel. If you mean "you", you should write it - thank you! Second: The only one not listening are apparently you. Multiple times you have been asked to measure your avatar size via a prim and check the prim's size in the edit window. And then compare what is reported in Firestorm and the Linden Lab viewer. Don't want to do it? Well, can't help you then! Third: If you think that showing the actual, correct avatar size related to prim sizes is wrong and a "megabug", but showing a size that doesn't match the size of a prim defined with the size as reported, then... well... guess can't help you either. But I can guarantee you the "megabug" will be in the next Firestorm release as well. And that's it from me about this topic. 😁
  5. You already said that in the original post, but failed to provide photos Beq asked for, showing where the size of your avatar reported in Firestorm differs from the size of a prim matching your avatar height. I tested various shapes myself now and compared the size reported in Firestorm with the size of a prim matching the avatar sizes and they pretty much match. In contrast, the size reported in the official viewer is way off compared to the size of the prim. Of course I tested Cool VL as well and if I take the size reported for shoe height into account as well, the reported height pretty much matches the height reported in Firestorm - guess you'll also have a "megabug" in your viewer as well, Henri - according to the OP. 😉
  6. Being 15 years in SL, you should know that avatar height in SL is relative and a lot of shape "creators" create giants nevertheless, seeing those 2.2m+ giants regularly showing up at events - which can't be caused by Firestorm if it calculates the size too small. 😉 In the end it boils down to creating a proportional correct shape anyway - having a correct height display in the viewer won't prevent stilt-legs, t-rex / orangutan arms or shrunken heads... 😁 And as I said: Once confirmed that the fix for LL's original flawed height calculation is obsolete and now actually results in a wrong height calculation, it will be removed to match LL's calculation.
  7. That is interesting, because when LL first came up with the height calculation, it was off by 19.5cm. Apparently it got corrected at some point (bento maybe???). Gonna confirm and fix for next release if necessary - no need to run a viewer that hasn't seen a release in a year...
  8. Havok is a physics engine and has nothing to do with OpenGL. The fact that SL is CPU bound is that purely the age of the application itself and LL cared little to nothing in modernizing their OpenGL implementation, probably so people on 10+ year old toasters could still run SL.
  9. As of now, this is true. This might change with the soon to be expected release of the performance viewer. Mostly true, but see answer for first question: The performance viewer and ongoing work based on it will make more use of threading, so having more cores will be an advantage here. That's actually true and not just related to SL, since this allows dual channeling which is faster than single channel memory access.
  10. It's a comprehensive list of changes/fixes/improvements in relation to the current default release viewer.
  11. I already merged all that stuff - and took it all back out, because we think living at high speeds is not good for your health and 5 FPS is enough when already visiting "wild" events.
  12. Sounds more like finally time to say goodbye to baked textures and say hello to advanced lighting with proper shadows! BTW: Most new residents are running shared environment settings instead of a fixed midday setting.
  13. Dedicated GPU memory is texture memory + FBO + VBO etc - texture memory is still capped at 512MB 😂
  14. Ohohoh, listen listen listen! It's not our fault Mac and Apple stuff in general is expensive broken crap! Are you mad yet? 😂
  15. Suddenly trying to derail? Well then... How about this: Works fine here - go fix your crappy hardware! 😛
  16. If that is officially available where you got your numbers from, then just link it here, otherwise your numbers are just pure speculation.
  17. Which is perfectly fine since there are about 95% Windows users in SL. 😁 Also, I would like to see the source where you got the 1.3 billion users from. Here they mention 1.8 billion devices (https://www.macrumors.com/2022/01/27/apple-1-8-billion-active-devices-worldwide/ ). Given that Apple cultists often own more than one device, please show me a source of the active user base.
  18. Sounds to me like it's outdated and needs to be deprecated according to Apple's own logic - and to the logic of some people here in the forum. 😋 My hobby is to teach everyone what an overhyped piece of crap and rip-off Apple is! 😊
  19. Nice you mention the Voodoo cards that were basically the break-through of OpenGL (aside they also had their own Glide API) with the release of glQuake back then... 😊 Shame you can't do a 1-on-1 transition from the OpenGL implementation in the current viewer to Vulkan, only to show those "Vulkan makes SL faster fetishists" that a simple switch to Vulkan does absolutely nothing. And by the time you refactored the whole viewer to use current OpenGL techniques, you will notice the "OpenGL is just bad" argument is pretty much bollocks. 😋 It's called "deprecated"... It's called "backwards compatibility" - something that makes greedy Apple managers shudder and sweat from fear, meaning they can't squeeze more money out of their "cultist" users by artificially breaking stuff, calling it "innovation" and making them pay for the same thing over and over again. OpenGL performance has ever been bad on AMD/ATI cards, long before they developed Mantle. What's the threshold for "being on a dead end"? Just curious, since the latest Metal API update was almost 3 years ago - might be on a dead end very soon, too! And better not check, when the last Collada update was.
  20. Does that make it automatically bad? The problem is not that the OpenGL spec hasn't been updated since 2017, but the fact the viewer implementation is still vastly depending on tech from 2005. And it's still Apple to blame for not releasing a proper compatibility profile back with OpenGL 3.2, effectively locking the viewer in OpenGL 2.1 since you can't mix newer features with old fixed-function pipeline stuff still in the viewer (which will hopefully change with the performance viewer, but I still found some old ff stuff in there...) You mean the rather 5% of people being unable to run SL anymore in contrast to the larger number of people with hardware that doesn't support something like Vulkan? Take a guess on the business decision. (On a sidenote: If you are a big creator in SL making lots of money and making a living from it, you'd probably get a Windows PC if Mac support is gone rather than saying byebye and going out of business.) Of course they don't wanna share anything - it's Apple! The only thing Apple is sharing is their walled garden and let you pay for that "experience"! Can't even exchange an SSD anymore so they can charge you ridiculous prices for standard hardware labeled as "Apple certified quality product", made in the same fab on the same machines they produced the same thing for the Windows/Linux PC plebs in the other shift!
  21. It's only deprecated by Apple trying to convince people it is "outdated technology" while in reality they were too lazy to fix their broken implementation of it and push everyone into using their Metal API for advanced vendor lock-in effects and squeezing more money out of their users.
×
×
  • Create New...