Tech Support Blog

  • Subscribe to our RSS feed.
  • Twitter
  • StumbleUpon
  • Reddit
  • Facebook
  • Digg

Thursday, 29 November 2007

How Framerate Dies - Glitching

Posted on 20:26 by Unknown
Back in the good old days (that would be X-Plane 6), X-Plane's framerate would suffer in two ways as you cranked up the rendering options:
  1. For most features (more visibility, more autogen) as the CPU and GPU became more heavily loaded, the framerate would gradually decrease.
  2. If you ran out of VRAM (that is, the working set of textures needed per frame was more than your card's VRAM) framerate would really die fast - think 2 fps.
The reason for this second behavior was that computers couldn't shuffle textures between main memory and VRAM fast enough to render a frame in a 30th of a second.

As computers have gotten faster, this second behavior has gone away - modern cards, with fast PCIe 16x busses, can transfer textures from system memory to VRAM pretty fast - fast enough to have the working set be (slightly) larger than VRAM and still fly. So as texture memory increases, framerate decreases more gradually.

However, a new behavior has emerged: "glitching". You may have noticed that when you've got your computer set to the ragged edge of the rendering settings, as you turn the camera, the framerate will stutter for a few frames, then return to a relatively high rate (40-50 fps).

What's happening is: the working set of textures and geometry needed by X-Plane just barely fit in VRAM. But when you turn your head, a different set of textures and geometry are needed. While the card sorts out what is needed and what isn't, it spends some time needlessly shuffling textures, and eventually reaches stability, with only what's needed in VRAM, and framerate stays high.

Glitching has emerged as a mode of performance degredation because over time we've cut down the amount of "stuff" (textures and geometry) x-plane needs to draw a frame to only what's really absolutely needed. This means there is less intersection between the working set in one view and another, and it also means you can get closer to the edge of your hardware.

So my view on glitching is basically "too bad". If the working set weren't as carefully trimmed, you wouldn't have glitches, but the framerate would be entirely low, not entirely high. The only solution is to turn down settings that increase the working set (object density, world LOD, tex res, forests...) until the computer can run without glitches.

An even stranger variant of this: users sometimes report framerate getting "stuck" at 19 fps and then coming back when they change apps. The problem is that the driver doesn't know exactly what the best order of textures to keep in VRAM vs. shuffle is...as the view changes, sometimes the driver ends up with a non-optimal decision about what stays in VRAM and what goes, causing framerate to drop. Changing which app is in the foreground fixes this by temporarily pushing a lot of items out of VRAM, at which point the driver makes a different decision by luck.

Again, the solution is simple: turn down rendering settings to get the working set smaller than VRAM.

Basically, if the working set is smaller than your amount of VRAM, you should have even framerates, proportional to rendering load.

If the working set is greater than VRAM, the driver may find an optimal way to shuffle things and only decrease fps a little, or it may find a non-optimal way to shuffle and you'll get terrible fps, and that shuffle can change on the fly, causing framerate to fluctuate all over the place.
Email ThisBlogThis!Share to XShare to FacebookShare to Pinterest
Posted in inside x-plane, performance | No comments
Newer Post Older Post Home

0 comments:

Post a Comment

Subscribe to: Post Comments (Atom)

Popular Posts

  • Developer Hardware
    So...just how awesome is my main development machine? Not that awesome. Periodically users ask me what my setup is. Usually the user wants...
  • That's one biiiiig polygon
    Something I'm seeing now that WED is in beta: airport layouts with the entire taxiway structure made from one really complex polygon. I...
  • Caught With My Pants Down
    My friends say I have become a technological curmudgeon...whenever a new gadget or device or operating system comes out, I just grumble abo...
  • Who Am I?
    This week we've seen an increase in questions from new users, potential customers (both in the consumer and professional spaces) and thi...
  • Mirrored Normal Maps
    Normal maps in X-Plane 940 have a funny property: if you flip the normal map horizontally or vertically, the bumps change direction. Things...
  • What is a panel region?
    X-Plane 9 introduces a new OBJ feature: panel regions. The basic idea is this: In X-Plane 8 you could use the 2-d panel as a texture in you...
  • The Future of WED
    WED 1.0 has gone RC . The on ly change from beta 5 is that I have the latest manual changes from Tom (including Cormac's illustrations ...
  • The 3-d Panel Is Not Always Necessary
    There is no need to use the 3-d panel if you only want 3-d cockpit. That might be the most counter-intuitive statement in the entire univers...
  • OS X 10.6.3 Performance
    OS X 10.6.3 is out. Besides adding a bunch of OpenGL extensions*, it looks like vertex performance is improved on nVidia hardware. My quic...
  • Bad Alloc Crashes in 920 - Bad Timing
    I just received a series of reports today that certain converted scenery will cause X-Plane to crash with a "bad alloc" error. Ba...

Categories

  • absurdly cute
  • Air Traffic Control
  • aircraft
  • Android
  • animation
  • announce
  • cockpits
  • documentation
  • drivers
  • file formats
  • global scenery
  • Goofy Screenshots
  • hacks
  • hardware
  • hobbies
  • inside x-plane
  • installer
  • ipad
  • iphone
  • legal
  • localization
  • modeling
  • off topic
  • palm pre
  • panels
  • performance
  • plugins
  • political
  • scenery system
  • tools
  • X-Plane 10
  • XSquawkBox

Blog Archive

  • ►  2011 (12)
    • ►  February (1)
    • ►  January (11)
  • ►  2010 (111)
    • ►  December (4)
    • ►  November (4)
    • ►  October (10)
    • ►  September (9)
    • ►  August (12)
    • ►  July (8)
    • ►  June (4)
    • ►  May (13)
    • ►  April (13)
    • ►  March (11)
    • ►  February (12)
    • ►  January (11)
  • ►  2009 (130)
    • ►  December (16)
    • ►  November (11)
    • ►  October (6)
    • ►  September (16)
    • ►  August (12)
    • ►  July (11)
    • ►  June (9)
    • ►  May (5)
    • ►  April (10)
    • ►  March (9)
    • ►  February (9)
    • ►  January (16)
  • ►  2008 (147)
    • ►  December (18)
    • ►  November (10)
    • ►  October (7)
    • ►  September (11)
    • ►  August (15)
    • ►  July (9)
    • ►  June (14)
    • ►  May (9)
    • ►  April (14)
    • ►  March (13)
    • ►  February (6)
    • ►  January (21)
  • ▼  2007 (100)
    • ►  December (17)
    • ▼  November (13)
      • How Framerate Dies - Glitching
      • Two Truths of Hardware
      • Obects and Liveries
      • A wide range of performance
      • Beta 3 - Don't Panic
      • Multi-Core and X-Plane 8 and 9
      • X-Plane 9 Performance - Let's Wait a Little
      • X-Plane 9 Is Here
      • ANIM_hide is not a framerate optimization
      • New SDK coming...
      • Add-on Forests...
      • The problem with KLGA
      • Taxiway Shapes
    • ►  October (13)
    • ►  September (9)
    • ►  August (17)
    • ►  July (7)
    • ►  June (4)
    • ►  May (6)
    • ►  April (9)
    • ►  March (2)
    • ►  February (3)
Powered by Blogger.

About Me

Unknown
View my complete profile