Jump to content
  • Advertisement
Sign in to follow this  
Vertex333

DX11 [DX11] Device recreation best practice

This topic is 2347 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hi all!

I have a 2D visualization app (D3D11 with Direct2D Interop) that may get the device reset/lost or whatever, so I have to recreate my device and resources. The app is critical and should run as long as possible or at least quit with an error message. The question here arises wheter I should try to recreate the device as long as possible (without any display to the user in the meantime), or should I stop after 1000 tries, or after a minute? What do other programs (I don't mean microsoft msdn samples, but real delivered/professional applications/games)?

I already do a software fallback to WARP if no HW is available or if recreation failed.

As far as I know a device reset could occur of the following resons (some of them may be incorrect and there are at least some more that I forgot): driver hangs & gets recovered (never had this/couldn't reproduce that), driver update/installation (couldn't reproduce that), adapter/device/driver gets removed (never had this), internal errors (had that but don't ask me what is needed to reproduce), insufficient memory (especially often with WARP, so not too hard to reproduce),...)

thx,
Vertex

Share this post


Link to post
Share on other sites
Advertisement
Wait one second before recreating the device, and if it fails wait ten seconds before trying again, and then just one try each minute, or something along those lines.

Share this post


Link to post
Share on other sites

Wait one second before recreating the device, and if it fails wait ten seconds before trying again, and then just one try each minute, or something along those lines.

Thx Erik!

So you would not end trying to recreate the device?
The problem with the app is that it is kind of a user interface to a "control", so it would be important to see graphics. A very long time without seeing anything shouldn't be that user friendly or in other words the user doesn't know what happens. On the one side we should not close our application, on the other side the user should not stay in front of our app without seeing anything + transistions/... that need device recreation should work!

The recreation is a very important part for our app (as described here). You can't totally prevent device losts/resets, so we should do the best. Hence I am very interested how other applications/games do that.

Share this post


Link to post
Share on other sites
You could display a message with GDI or a dialog box while waiting, saying that the device was lost and is waiting to be recreated, or even a button to attempt recreation right away. Perhaps store the times from creation until the device is removed, and if you find that it always gets removed in less than some number of minutes, stop trying to recreate it until the user asks for it.

Share this post


Link to post
Share on other sites

You could display a message with GDI or a dialog box while waiting, saying that the device was lost and is waiting to be recreated, or even a button to attempt recreation right away. Perhaps store the times from creation until the device is removed, and if you find that it always gets removed in less than some number of minutes, stop trying to recreate it until the user asks for it.

Thx again Erik!

That sounds already pretty nice, but it is not allowed to bring a message box in the application (except we close the app): nothing is allowed to interrupt user interaction and so on. The graphics is important, but the user should not have to interact for the graphics stuff. The application could simply run without any person beeing there for days (then the graphics don't need to be drawn, but we could not ask the user to do anything or block anything of the app or something like that).

The idea with the time measurement is good. Thx! Nevertheless, it is hard to set the times + no user interaction or user dependency.

Vertex

Share this post


Link to post
Share on other sites
If I were designing such a system, I would list each of the ways that a device could be 'lost', and then decide an appropriate measure for each of those cases while considering your provided constraints. According to this page, the WDDM driver model only allows the following reasons for a 'losing' a device:

"WDDM now provides a GPU memory manager and scheduler that allow multiple applications access to the GPU simultaneously. Because Direct3D applications no longer require exclusive access to the GPU it is possible to switch focus between applications with little penalty. Under WDDM Direct3D devices are only lost during driver upgrades, physical removal of the device, GPU reset and unexpected errors."

So each of these cases has its own semantic meaning, which means it should be handled specially:

  1. Driver Upgrades: This should only happen when someone is updating your control machine, which means you can assume someone is there. Displaying a non-interactive message window to inform them that the driver is being upgraded should be possible (if you can't already assume that your app will be shutdown for driver upgrades).
  2. Physical Removal of the Device: This follows the same logic as #1 - I think you can assume a technician is there and working on the machine if it is removed. This also assumes that the video card doesn't fall out of the machine, but that should be extremely rare...
  3. GPU Reset: I have seen this occur when there is a driver bug, or if a compute shader takes a very long time to execute. In this case, it is a systemic problem and will probably not go away on its own.
  4. Unexpected Errors: They are unexpected, so you don't know what to expect :)

Out of #1 and #2, I don't think you need to worry about them. They are extremely unlikely to occur unless someone is doing maintenance on your machine. #3 is likely to only occur in error or logical error situations, meaning you need to test against any possible erroneous inputs etc... Still, you won't know this one is coming until it hits you, but it should also only occur in special conditions. #4 is much the same as #3, they are conditions that are probably going to require a special situation to produce the problem, which I would assume is a non-repeating condition.

So my advice would be to retry to create the device every 5 seconds for up to 1 or 2 minutes. As soon as the first error occurs, I would open a GDI window with just a message indicating what is happening, and also to indicate the current status of the recreation (i.e. retrying in 5, 4, 3, 2, 1...). If it doesn't get corrected within that period of time, you need to assume something really bad has happened, and then switch to your backup with the WARP device. Please keep in mind that a WARP device doesn't necessarily provide 100% guarantee that the device will be created correctly either - i.e. if something happens to the dll that it resides in then you are stuck too... If this is a mission critical type of setup, I would provide even a GDI fall back for very basic, but functional, interface graphics.

And one last point - if you are running on a WDDM driver, I was not aware of the fact that you can run out of memory since the GPU memory is virtualized now. You would need to run out of memory on the whole machine before running out of GPU memory... How are you sure that you have an out of memory error?

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement
  • Advertisement
  • Popular Tags

  • Popular Now

  • Advertisement
  • Similar Content

    • By chiffre
      Introduction:
      In general my questions pertain to the differences between floating- and fixed-point data. Additionally I would like to understand when it can be advantageous to prefer fixed-point representation over floating-point representation in the context of vertex data and how the hardware deals with the different data-types. I believe I should be able to reduce the amount of data (bytes) necessary per vertex by choosing the most opportune representations for my vertex attributes. Thanks ahead of time if you, the reader, are considering the effort of reading this and helping me.
      I found an old topic that shows this is possible in principal, but I am not sure I understand what the pitfalls are when using fixed-point representation and whether there are any hardware-based performance advantages/disadvantages.
      (TLDR at bottom)
      The Actual Post:
      To my understanding HLSL/D3D11 offers not just the traditional floating point model in half-,single-, and double-precision, but also the fixed-point model in form of signed/unsigned normalized integers in 8-,10-,16-,24-, and 32-bit variants. Both models offer a finite sequence of "grid-points". The obvious difference between the two models is that the fixed-point model offers a constant spacing between values in the normalized range of [0,1] or [-1,1], while the floating point model allows for smaller "deltas" as you get closer to 0, and larger "deltas" the further you are away from 0.
      To add some context, let me define a struct as an example:
      struct VertexData { float[3] position; //3x32-bits float[2] texCoord; //2x32-bits float[3] normals; //3x32-bits } //Total of 32 bytes Every vertex gets a position, a coordinate on my texture, and a normal to do some light calculations. In this case we have 8x32=256bits per vertex. Since the texture coordinates lie in the interval [0,1] and the normal vector components are in the interval [-1,1] it would seem useful to use normalized representation as suggested in the topic linked at the top of the post. The texture coordinates might as well be represented in a fixed-point model, because it seems most useful to be able to sample the texture in a uniform manner, as the pixels don't get any "denser" as we get closer to 0. In other words the "delta" does not need to become any smaller as the texture coordinates approach (0,0). A similar argument can be made for the normal-vector, as a normal vector should be normalized anyway, and we want as many points as possible on the sphere around (0,0,0) with a radius of 1, and we don't care about precision around the origin. Even if we have large textures such as 4k by 4k (or the maximum allowed by D3D11, 16k by 16k) we only need as many grid-points on one axis, as there are pixels on one axis. An unsigned normalized 14 bit integer would be ideal, but because it is both unsupported and impractical, we will stick to an unsigned normalized 16 bit integer. The same type should take care of the normal vector coordinates, and might even be a bit overkill.
      struct VertexData { float[3] position; //3x32-bits uint16_t[2] texCoord; //2x16bits uint16_t[3] normals; //3x16bits } //Total of 22 bytes Seems like a good start, and we might even be able to take it further, but before we pursue that path, here is my first question: can the GPU even work with the data in this format, or is all I have accomplished minimizing CPU-side RAM usage? Does the GPU have to convert the texture coordinates back to a floating-point model when I hand them over to the sampler in my pixel shader? I have looked up the data types for HLSL and I am not sure I even comprehend how to declare the vertex input type in HLSL. Would the following work?
      struct VertexInputType { float3 pos; //this one is obvious unorm half2 tex; //half corresponds to a 16-bit float, so I assume this is wrong, but this the only 16-bit type I found on the linked MSDN site snorm half3 normal; //same as above } I assume this is possible somehow, as I have found input element formats such as: DXGI_FORMAT_R16G16B16A16_SNORM and DXGI_FORMAT_R16G16B16A16_UNORM (also available with a different number of components, as well as different component lengths). I might have to avoid 3-component vectors because there is no 3-component 16-bit input element format, but that is the least of my worries. The next question would be: what happens with my normals if I try to do lighting calculations with them in such a normalized-fixed-point format? Is there no issue as long as I take care not to mix floating- and fixed-point data? Or would that work as well? In general this gives rise to the question: how does the GPU handle fixed-point arithmetic? Is it the same as integer-arithmetic, and/or is it faster/slower than floating-point arithmetic?
      Assuming that we still have a valid and useful VertexData format, how far could I take this while remaining on the sensible side of what could be called optimization? Theoretically I could use the an input element format such as DXGI_FORMAT_R10G10B10A2_UNORM to pack my normal coordinates into a 10-bit fixed-point format, and my verticies (in object space) might even be representable in a 16-bit unsigned normalized fixed-point format. That way I could end up with something like the following struct:
      struct VertexData { uint16_t[3] pos; //3x16bits uint16_t[2] texCoord; //2x16bits uint32_t packedNormals; //10+10+10+2bits } //Total of 14 bytes Could I use a vertex structure like this without too much performance-loss on the GPU-side? If the GPU has to execute some sort of unpacking algorithm in the background I might as well let it be. In the end I have a functioning deferred renderer, but I would like to reduce the memory footprint of the huge amount of vertecies involved in rendering my landscape. 
      TLDR: I have a lot of vertices that I need to render and I want to reduce the RAM-usage without introducing crazy compression/decompression algorithms to the CPU or GPU. I am hoping to find a solution by involving fixed-point data-types, but I am not exactly sure how how that would work.
    • By cozzie
      Hi all,
      I was wondering it it matters in which order you draw 2D and 3D items, looking at the BeginDraw/EndDraw calls on a D2D rendertarget.
      The order in which you do the actual draw calls is clear, 3D first then 2D, means the 2D (DrawText in this case) is in front of the 3D scene.
      The question is mainly about when to call the BeginDraw and EndDraw.
      Note that I'm drawing D2D stuff through a DXGI surface linked to the 3D RT.
      Option 1:
      A - Begin frame, clear D3D RT
      B - Draw 3D
      C - BeginDraw D2D RT
      D - Draw 2D
      E - EndDraw D2D RT
      F - Present
      Option 2:
      A - Begin frame, clear D3D RT + BeginDraw D2D RT
      B - Draw 3D
      C - Draw 2D
      D - EndDraw D2D RT
      E- Present
      Would there be a difference (performance/issue?) in using option 2? (versus 1)
      Any input is appreciated.
    • By Sebastian Werema
      Do you know any papers that cover custom data structures like lists or binary trees implemented in hlsl without CUDA that work perfectly fine no matter how many threads try to use them at any given time?
    • By cozzie
      Hi all,
      Last week I noticed that when I run my test application(s) in Renderdoc, it crashes when it enable my code that uses D2D/DirectWrite. In Visual Studio no issues occur (debug or release), but when I run the same executable in Renderdoc, it crashes somehow (assert of D2D rendertarget or without any information). Before I spend hours on debugging/ figuring it out, does someone have experience with this symptom and/or know if Renderdoc has known issues with D2D? (if so, that would be bad news for debugging my application in the future );
      I can also post some more information on what happens, code and which code commented out, eliminates the problems (when running in RenderDoc).
      Any input is appreciated.
    • By lonewolff
      Hi Guys,
      I understand how to create input layouts etc... But I am wondering is it at all possible to derive an input layout from a shader and create the input layout directly from this? (Rather than manually specifying the input layout format?)
      Thanks in advance :)
       
  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

Participate in the game development conversation and more when you create an account on GameDev.net!

Sign me up!