Brett and a few other users pointed out to me that the nVidida ForceWare 169.21 release appears to hose video for X-Plane. If you have an nVidia card, don’t update, or you may want to back up a version.
This kind of thing has happened in the past – hopefully a revision will come out fairly soon. (But I do not have contact with the nvidia driver team for Windows on this…)
Also some users are seeing corrupt startup screens with ATI hardware – apparently some configs don’t react well to us turning vsync off. Not quite sure what we’ll do about that yet, but hopefully a fix will be in the next few betas.
We’ve been preaching “one big texture, not lots of little textures” for a while now, and generally speaking, packing a lot of art into one big texture makes life eaiser for X-Plane, because it can draw more triangles at once before it has to tell the card to change what it’s doing. Inside the company we call this the “crayon rule“.
Now the total set of geometry and textures that X-Plane needs to use for one frame is the “working set” – you can think of it as the crayons that you keep out of the box because you need them all the time. And as I said before, if the working set becomes too big, your framerate dies.
Now with large panels we’re seeing a new phenomenon, one of the first cases where the crayon rule might not be true. The reason is due to working set.
When you make an airplane with a large panel in version 9, you can either use ATTR_cockpit, which lets you use the entire panel as a texture, or you can use ATTR_cockpit_region, which will let you use several parts of the panel. Each ATTR_cockpit_region is a texture change, so that’s more crayons. And yet ATTR_cockpit_region is usually faster.
The reason is two-fold:
- You can often use cockpit regions that don’t cover the entire cockpit texture. Large panels are rounded up to 2048 if the are larger than 1024 in any dimension, so the “wasted space” in a 1600×1600 panel is actually quite huge. If you can get away with some smaller regions, your total panel texture area is smaller because there isn’t wasted space due to this rounding, and you can also skip things like Windows. Prepping the panel texure takes time, and it’s done once for lit and once for non-it elements, so it adds up!
- It turns out there are two categories of textures that contribute to the working set: static texures and dynamic ones, and their impact on VRAM is very different. Dynamic textures are much more expensive. The panel texture is dynamic and it’s uncompressed, so it really costs a fortune. (32 MB of VRAM for 1600×1600. That’s not a lot for a static texture but for a dynamic one that’ll kill you.)
Here’s the details on dynamic vs static textures: the OpenGL driver keeps a backup copy of a texture in main memory, so that if it has to purge VRAM (to make room for more stuff) it still has the texture. As it “swaps” textures, the process is to simply send textures as needed from main memory to VRAM. No big deal.
But with a dynamic texture, the texture has been modified in VRAM! So the copy in system memory is old and stale. The graphics card thus must send the texure back to main memory, consuming twice as much bus bandwidth as normal. (To free 16 MB of VRAM and refill it takes 32 MB of transfer, 16 MB to copy the old texture back to system RAM and another 16 to send the new textures to VRAM.) On non-PCIe cards, this back-transfer might be at 1/8th the speed of the transfer to the card, so this is even worse on AGP cards.
Thus the driver does its best to not throw out dynamic textures. And this is why the panel texture is so expensive. That P180 will cause X-Plane to make two 16-MB dynamic texures, and those textures will cause 32 MB of VRAM to basically be off the table. That’s less space for the other textures to swap in and out of. This kind of “permanent allocation” makes the VRAM budget tighter for all other drawing operations.
Given the right combination of large panels, large res, pixel shader effects (which make more dynamic textures), clouds, and FSAA, you can easily get even a 256 MB card to a state where the free space into which static textures are shuffled becomes horribly small, and the framerate just dies.
So the moral of the story is: yes, it can be worth 4 crayons (using panel regions) to avoid the huge cost of dynamic textures from large panels.
As to static textures (regular DDS files) that are 2048×2048 – the jury is still out but my guess is they don’t represent a huge performance problem. As one user pointed out to me, they’re only 2 MB when compressed (maybe more with alpha) so they’re not insanely huge, and they can be swapped out.
Lori and I are about to leave for a New Years Eve ski trip, but before I shut down the laptop for the last time in 2007 I wanted to say: Happy New Years to everyone in the X-Plane community. I had a lot of fun working on X-Plane in 07 and hopefully the sim brought you enjoyment too. I think 2008 is going to be very exciting – version 9 plants the seeds for a lot of interesting new possibilities during the version run.
When I get back I’ll post a bit on panel regions, for which I have the first performance numbers, as well as some of the strange effects FSAA has on the sim. We should have a progress report on Linux soon too.
See you next year!
Sometimes we find that users machines cannot run without hiding OpenGL driver features from X-Plane. That is, the computer says it can support VBOs, but when the sim asks for a VBO, something really bad happens. X-Plane (since mid-version 8) accepts a series of command-line options that cause the sim to ignore the given feature.
These kind of bugs come and go as drivers are updated, the sim changes which technology it uses, and hardware cycles through the user base. The biggest one we’re seeing right now is that the new iMacs show runway lights as white squares unless the sim is run with the –no_sprites option.
We’re trying a new way to address these problems. In the past we would give users the command-line option; now we are building double-clickable script files that launch the sim with the appropriate options. Theoretically this…
- Is less error prone for users.
- Is quicker for users who may have to use the command-line option every time they launch the sim until a driver update becomes available.
- Is quicker for us (in that we spend less time mailing out instructions and helping users who are unfamiliar with a command-line environment).
We did consider some other options, but this seemed like the least evil. The runners-up:
- Just turning off the hardware feature permanently. We ruled this out because the performance hit would be significant and affect all users.
- Attempt to identify and auto-turn-off known bad options. We do this in some cases, but it requires changes to the sim code, so this does no good if a new video card comes out after a given x-plane release and introduces new problems. so it’s not a total solution. Also since a bug might be resolved in-field, if X-Plane auto-avoids certain configurations we have to patch the sim once the configuration starts to work again.
- Provide a user interface to turn these options off inside the sim. This was ruled out for two reasons: first, some of these options cause the sim to crash before a user can ever get to the rendering settings. Second, turning off these kinds of options can really kill performance, so leaving the options in sight produces a whole new tech support option. (A user tells us their framerate is awful at the lowest settings…perhaps they turned off the hardware acceleration options in an attempt to get the lowest settings…etc.)
We’ll see how well this approach works. So far it seems to be working better than handing out command-line options.
(Of course if we can address the problem by working around it with a change to our code, we almost always choose this option.)
In X-Plane 8 the global scenery was installed in the “default scenery” folder, which normally contains most of the built-in artwork to make X-Plane (and the global scenery) work.
In the X-Plane 9 beta it is installed in the custom scenery folder. But the sim recognizes it by name and makes sure custom scenery is higher priority than global scenery. This is imperfect but works. In the future we may do something different. Survival tips:
- The log file tells the priority of all scenery packages.
- Don’t rename or move the global scenery – leave it where it was installed.
- Custom scenery will always trump global scenery.
In X-Plane 9, road networks can be in DSF overlays. And in X-Plane 8 or 9, road types can be customized using a road.net file.
Within the road definition file, each road type can contain:
- “Segments” (textured quads that make the roads)
- Objects (placed along the road)
- Wires (for powerlines)
- Car lanes (that vehicles drive along)
The interesting thing is: none of these types are mandatory. So you can use any of these features without the others.
“Cars” are just objects too. So one way to make a vehicle drive around an airport would be:
- Create a road definition file with a new road type. That road type only has a car lane which references your vehicles OBJ model.
- Make an overlay DSF with this new road type, drawing the path the vehicle OBJ will randomly take.
- Package the overlay, .net file, and OBJ all together and you’ve got randomly generated vehicles moving around.
As Marginal pointed out on the .org forums, roads are currently not draped, so using them in overlays is a total PITA. This is something I hope to address in the 9.1 time-frame…draped roads are an important feature, but a little bit too invasive for the v9 betas.
I found the root cause of another NVidia specific bug, and once again it’s my own stupid code. If you Google for driver bugs, you’ll find plenty of grumpy developers ranting about how card X does this wrong thing and card Y does that wrong thing…I figure it’s only fair to follow up and say “yep, that one was mine.”
Like the previous nVidia-only crash, this was a case where X-Plane was always doing something wrong, but only some drivers had problems with the behavior. So the crash was NVidia-specific, but X-Plane caused.
I believe that this bug was manfiesting itself either as a message that “scenery shift took more than 30 seconds” or some kind of crash. One of the problems was that the diagnostics for this particular bit of code were really bad. So we’ve improved things a bunch…
- There is more careful error checking during scenery shift, and those error messages are reported.
- If the sim does crash, some new code will output a crash log on Windows that helps us isolate what actually happened.
Beta 12 will be out soon with the fix that caused problems on NV hardware as well as the improved diagnostics. So you may find that the sim just works better, but if it does still crash or report errors, please tell us – now we’ll have log files that will let us diagnose the problem a lot faster!
I have spent perhaps the last two weeks tracking down driver-related problems. But the term “driver bug” is heavily overused (blog around and you’ll see that many OpenGL developers get frustrated…) A few examples of the gray areas:
- Sometimes there will be a bug in application code that shows up only on certain hardware. Drivers are concerned with making video cards go fast, not spanking programmers who don’t know what they’re doing. This is exactly what happened to me in beta 2 with the crash-on-nVidia-with-C172 bug. This was just plain broken code in X-Plane, but for some reason the ATI drivers didn’t have a problem with it (probably because they were performing an optimization which let them ignore the bogus call I made). NVidia specific, but not NVidia’s fault!
- On OS X with the Radeon 9600XT, runway lights don’t show up. Adding an extra line to the pixel shaders “fixes” the problem. I believe the problem is in the driver (or the shader compiler more specifically) but by changing the code to the shader we work around the problem. A change in X-Plane addresses the problem, but not X-Plane’s fault!
- The “36061” errors that some users have been seeing turn out to be because (through a very convoluted chain of events) X-Plane was asking the video card to operate in a mode that it could not operate in. Turns out this can be fixed by changing X-Plane’s code (fix will be in beta 12) or by getting new drivers. This one wasn’t really a driver bug – more that the drivers were limited in a way X-Plane did not expect. (Our fault for being picky!)
The situation is fundamentally tricky – games are integrators of other people’s technology – as such, we get blamed by the end user for a fault anywhere in the system. At the same time, it’s way too easy to turn around and blame the part supplier, and unfair when the source of the bug hasn’t been identified.
I am looking now at problems on Windows with dual core machines and nVidia cards. The problem goes away both by changing a registry setting that affects the driver and by changing X-Plane’s code. So I think it’s too soon to tell on this one.
I’ve spent perhaps the last two weeks almost entirely looking at driver-related graphics bugs. I should note that these are often not driver bugs, and the issue of whether a bug is in the sim or a driver is a lot stickier than you’d imagine, but I’ll blog about that later.
In the meantime, where do we stand on scenery development tools? Here’s a rough summary.
General Strategy
The scenery tools are not targeted to X-Plane 8 or 9 – they’ll export content for both (but if you use new V9-only features, your content will be v9 dependent).
There’s been one structural change in v9 that’s worth mentioning: version 9 allows road networks in DSF overlays. This cleans up the scenery model a lot…roads are more like objects and forests than like the base mesh, so now all of the 3-d elements can be in an overlay. This is particularly useful because you might want to adjust the road and objects nearby all at once.
Given that we have no road editing tools right now, I think there is one case where the tools will be v9-centric: I do not think I will ever provide a road-editing tools for DSF base meshes with roads. I do think I will add road editing to the list of overlay features that will eventually end up in WED.
This means that if you want to customize roads, you’ll have to do it in version 9. This is an inevitable outcome of the scenery system structure…replacing the entire base mesh to edit the roads is not a great idea.
WED
WED is in beta 4, and I think it will go final as soon as I have time to make it official. At this point most of the open bugs I get are usability issues and enhancement requests, not real bugs. We’ll get more usability features into WED in the future.
Please don’t send me WED bugs now (or attach them as blog comments). I’ll put out a formal last-call for WED bugs when I have time to dig in.
The future plan for WED is to add DSF overlay editing with all legal overlay types (which is now everything except beaches and base mesh).
AC3D
The AC3D export plugin already supports new v9 OBJ features – key frames and panel regions. These tools are available now!
Tools
The “tools” are the set of other scenery apps: DSF2Text, ObjConverter and ObjView. We’ll be adding two new tools soon:
- DDSTool is a PNG->DDS converter. There are already other programs (many free) for both platforms that can do this conversion, but we wanted something easy and drag & drop when we were doing the in-house conversion, so I coded DDSTool (based on the Squish library). We’ll make it available to authors.
- MeshTool is a base-mesh generation program. We’re still working out the details on this, but basically it will generate DSF base meshes from raw DEM data and polygonal coastlines.
I am restructuring the tools to be entirely command-line-based (except for ObjView) with a single “drag & drop” UI that wraps all of the tools. So if you use the mouse, you’ll use the one drag & drop application to do all your DSF2Text converions, 3DS->OBJ conversions, PNG->DDS conversion, etc. This is mainly to cut down the number of apps I have to build, but it should also declutter workflow a bit.
MeshTool
I’ll post more on MeshTool once we have more details, but here are the main points:
- MeshTool will not have a UI. It will take text files and process them. It is designed to provide a building block on which people can write more advanced tools. In other words, MeshTool gives other programmers our mesh algorithms just like DSF2Text gives other programmers our DSF-packing algorithms.
- MeshTool will not be an editor. It will create NEW meshes from RAW DATA, not from previous meshes. There are a lot of reasons why I do not want the mesh-creation workflow to be “take the global scenery and hack it”. So MeshTool lets you build your own.
- This means you will need your own elevation and coastline data for MeshTool! You can’t use what shipped in the global scenery! However, since the data that ships the global scenery is freely downloadable and widely available, I don’t think this is a problem.
- MeshTool will be able to “burn in” orthophotos into a mesh. This is probably its most important feature. Right now the only way to make orthophoto-based scenery is to use overlay polygons, but the performance of .pol files with a lot of scenery is really poor under version 9. MeshTool will let people build photo-sceneries that take full advantage of the framerate that X-Plane can deliver!
MeshTool should address some of the common issues I’m hearing with orthophoto sceneries (slow performance with a ton of .pol files, no ability to fix the coastlines or remove underlying base terrain, no ability to fix elevation).
MeshTool will create new default landuse terrain where there are no orthophotos or water. This is important because the elevation data can be customized. If you take an existing DSF and somehow edit the elevation points (for example, use DSF2Text and just start editing) and raise a grassy field, you’ll have grass going up a cliff. But if you pass the raw elevation data to MeshTool, it will build a rocky cliff over steep elevation, creating more plausible scenery.
MeshTool should be in the post 9.0 final time-frame, as will all the other tool reposts.
Posted in Tools
by
Ben Supnik |
At this point my in-box has approximately 180 emails from the month of December regarding X-Plane 9. So while I appreciate all of the feedback we’ve gotten (bug reports, performance, etc.) it’s going to take a while to reply. If you haven’t heard from me, don’t panic! I hope to answer a whole pile of emails next week.
In the meantime, I’ve been working on improved crash logging on Windows. Right now when we have a crash on Windows, all we know is that (1) X-Plane crashed and (2) what DLL we crashed in (which is always us or the video driver – not useful).
Coming soon, X-Plane will catch the fatal crash, examine memory to see what was going on, examine its own EXE to figure out the names of the functions going on, and output it all to a crash log that users can send us to get a much clearer picture of what’s going on. This information is called a “backtrace” – we’ve had it for the Mac or a while (OS X provides back-traces automatically with a crash) and it’s really useful.*
So my top priority is all of the users who are seeing problems during scenery load, and a new build with a back-trace should help to reveal what’s really going on.
I’m also working on putting additional timing and performance information into the sim so we can learn more about why some users have poor performance. So far here’s what I’m seeing:
- 8800 users seem to have great performance. If you have this card and don’t have good fps, adjust your x-plane settings and NV control panel settings – this card can bring it.
- 8600 users sometimes have performance problems – not sure why.
- Older nvidia GPUs (7600, 6800) sometimes have performance problems with the new eye-candy features – I am investigating.
- The pixel shaders seem to slow down the new HD2x00 Radeons a lot more than expected…I still need to investigate this. This is the most surprising datapoint – the X1600 does very well, so I would expect newer GPUs to at least have that level of performance. I think this is something we might be able to address.
However not all of the reports are consistent, so I think it’s too soon to make some calls on recommended hardware. The only thing that’s clear is that most 8800 useres who we do careful perf experiments with end up with huge framerates.
* Microsoft provides some back-trace facilities, but since we don’t use their compiler tools, we had to roll our own.
