Jump to content
  • Advertisement
Sign in to follow this  

DX11 [HLSL/DX11] Tessendorf waves issue

This topic is 2375 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

Hello,

I'm trying to generate deep ocean waves according to the Tessendorf model (Phillips spectrum + FFT). So I've got two compute shaders, one that runs at the start of my program and generates the base heightfield in frequency space "h0", and another which runs every frame and generates the frequency heightfield according to the current time ("ht"), and performs a two-dimensional FFT over this field to produce the final heightfield containing the wave heights at a certain time t.

My heightfield is 512x512 in size, so my compute shaders run on 512 threads, and each loops over the row/column that it's currently processing; so the first compute shader, each thread processes one row, and in the second compute shader, at the first pass each thread processes a row, and then processes a column.

But it doesn't work. This is the output I'm getting at the various stages of the program (mapped onto a vertex grid by displacing the vertex heights):

- Generation of "h0"

2cei4xx.png

The first FFT pass (on the rows of "ht"):

1ytthl.png

And the final heightfield after both FFT passes is.. well.. apparently it's just a bunch of NaN's. I don't know what I'm doing wrong, I'm getting no debug warnings from DX11 so I suppose it's an algorithm problem. But really this is difficult, because I have no reference at all to work with except a paper of which I don't understand half the math and a piece of CUDA code for generating "h0" and "ht" (CUDA has an inbuilt FFT so I can't see how they do it). And using CUDA is not an option because I have an ATI card (this also means I can't run the reference code and compare outputs at various stages). I tried several different FFT implementations and they all produce the same result... I've been spending days on this and I'm getting desperate. Could anyone help me? Here is the shader code:

Generation of h0:http://pastebin.com/N3SXHeNj

FFT with ht calculation: http://pastebin.com/Br2qjjB3

My global.hlsl file (contains utility functions and constants): http://pastebin.com/tt5qmvp5

And the "complex.fxh" file is just complex arithmetic (which works perfectly because I use it in my fractal renderer too).


The FFT stage works as follows: set the h0 heightfield as a shader resource (texture2D), then perform the "ht" calculations along with the FFT row pass, and output the result to the output buffer. Then, copy the output buffer to the input buffer, and perform the FFT column pass on the input (for each pass the FFTPASS constant buffer member is updated).

Does anyone know where the problem is?? :( Please tell me if you need more information.. I'm already not that good in higher-level mathematics so all these complex equations are really confusing. I really tried to understand them and implement them the best I could but I'm at my wit's end here.

Thanks!!

Share this post


Link to post
Share on other sites
Advertisement
I'm also getting artifacts where certain rows/columns of the heightfield will seem to change wildly from frame to frame (but only very select ones, and always the same ones), as if there was uninitialized memory somewhere (if that makes sense).

Share this post


Link to post
Share on other sites
Havnt looked at your code in detail.


But I wrote something equivelant using OpenCL, the first thing to do is verify each part individual. Start with the FFT routine and test some common functions and compare against Maxima (or other known working FFT). You could also try implimenting parts on the CPU(in fact IIRC my initial generation is always done on the CPU) and comparing the results to what you expect.



David

Share this post


Link to post
Share on other sites
Debugging compute shaders can be painful, but it's possible. For AMD GPU's you can use GPUPerfStudio, although I'll warn you that it's very clunky. Another low-tech way of debugging/verifying results is to simply copy your compute shader resolutions to a staging buffer/texture so that you can read it back on the CPU. It will kill your performance, but it works.

Share this post


Link to post
Share on other sites
Right, well I just found out something... I managed to get the FFT more or less working (it now looks somewhat like water, just with some high frequency components at the edges, dunno why, will debug, probably a texture addressing mistake - I'll try porting the FFT to the CPU and see what the result is), and so I'm left with this weird vertex stuttering issue I described in my last post. However I discovered something: when I run my code I get the stuttering, however if prior to starting the exact same code I launch some GPU-intensive program like a benchmark (furmark, etc..), then the stuttering disappears!!

Could this be a driver bug, or is my card fried or something? It does tend to fail on some particular applications and games but overall this is the first time I see it artifact like this. What's going on? No wonder I can't produce working stuff if I can't trust that my hardware respects logic, lol.




I tried the reference rasterizer and it seems to not artifact in REF mode but it's so damn slow I can't really be sure.




EDIT: thoroughly tested RAM/VRAM with memtest86 and memtestCL... no issues. I'm going with driver bug of some sorts :angry:

Share this post


Link to post
Share on other sites
Ok I ported the code to the CPU and it turns out I needed to apply sign correction to the real parts of the resulting FFT (for some reason). Going to port it back to the GPU and see if it works now. Once I get that working I'll be able to focus on calculating the normal and "chopiness" FFT's.

CPU heightmap result (the rasterization code is kind of rushed but at least it shows the FFT is working):

t06uqw.png

Although that doesn't explain the GPU glitch I was seeing before - I guess I can live with it but it's bugging me.

Share this post


Link to post
Share on other sites
Ok, it worked and I now get some nice FFT waves. But I'm completely stumped (once again :angry:) on how to get the analytic normal vectors. I don't quite understand the formula given in the PDF (something with multiplying the fourier amplitudes of each point by "ik" where k is the frequency vector for that point). I have no idea how it's possible to get a 3D normal vector out of a single 1D value obtained from the normal FFT... all the implementations I've found on the net just resort to finite differences but I want to use the analytical normal because it should look better right? Really the PDF doesn't elaborate at all on how to obtain the normal, it just goes "the slope vector can be obtained like this [insert large formula]", I guess it must be obvious or something but I really don't get it. Could someone shed some light on how to do this please? I would be very grateful!




PS: the paper I'm talking about is this one: graphics.ucsd.edu/courses/rendering/2005/jdewall/tessendorf.pdf

Share this post


Link to post
Share on other sites
Update: found this online after a last ditch effort for finding something on google:

http://www.keithlantz.net/2011/10/ocean-simulation-part-one-using-the-discrete-fourier-transform/

(there is a second part too which is important as well)

It covers pretty much everything I was confused about and actually contains useable information for people like me, perhaps others will find it useful. :)

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!