Jump to content
  • Advertisement
Sign in to follow this  
kRogue

OpenGL optimal VBO usage?

This topic is 3661 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 all, I have read on posts here at gamedev.net, that one gets better performance out of VBO's if one creates VBO's of about 1-4MB in size, thus rather than have one VBO per mesh, one combines several meshes data into 1 VBO. In my situation I am also talking about static VBO's, i.e. those VBO's fro which the data does not change. Now my question: what kind of performance increase have people seen in combining into a few VBO's? I ask because at developer.nvidia.com, there is no mention of an optimal VBO size, or more vs fewer VBO's are better. There is mention to make VBO's before textures, that do use GL_UNSIGNED_SHORT for indexing and that binding VBO's is cheap, but glVertexPointer() and like functions are more expensive (so in that case one can easily argue to combine geometric information of a fixed mesh into one VBO and use an interleaved format, but to take advantage of that then one has to use the "opengl" names for vertex shader attributes, as I don't think there is an analogue for interleaved formats for named attributes, ala glVertexAttribPointer). At any rate to ask: 1) what kind of performance increase have people seen, if any, in using just a few large VBO's for many meshes rather than a simplistic 1 VBO per mesh behavior? 2) what kind of performance increase have people seen, if any, in using one VBO to hold different attribute and then using glInterleaved like calls, instead of having say one VBO per attribute and using multiple glVertexAttrib, or one VBO for all attributes and using glInterleaved like calls)? Note that implementing 2) is often quite easy, but implementing 1) is a real hassle.

Share this post


Link to post
Share on other sites
Advertisement
Sorry KRogue,

I can't answer any of your questions, because I am lacking the experience. The only thing I can do is to raise the general confusing :-P

I am struggling with the same problems and I am always happy to see that I am not alone =)

To 1)
My VBO-Performance is not too good, I tried a 1Mio triangle batch and performance went not well (without any shading).

So today I queried:
glGetIntegerv(GL_MAX_ELEMENTS_INDICES, &MaxElementsIndices);
glGetIntegerv(GL_MAX_ELEMENTS_VERTICES, &MaxElementsVertices);

And guess what, the results were:
4096 for both !!! *loooooll* ????

Is this for real ? What can I batch with THAT ? That sounds not so much like BATCHING in my ears, rather than SPLITTING !!??
Tell me that this is a bad joke. I am using a GF6800LE 128MB.

Btw. I need to batch up a lot of small primitives like rectangles, circles, polygons, so I have to fit many shapes (not meshes) into a single buffer for performance reasons. I know its a hassle. But then again with 4096 indices one can not batch too many primitives ;-)

To 2)
In theory I know cache utilization will be worse, but it won't kill you. I would suggest just do it, but again what do i know ?


My frustration with OpenGL + Documentation is constantly growing, I begin to believe that it may be possible that it is just too complicated for me. The unfortunate thing is I need to come up with a vbo-policy rather quickly for my work. Otherwise I would give it a rest.

I am tired, I will go now and implement *something* working...
Sorry for not giving the answer you are expecting - I have read the very same posts as you, but can't talk from experience :-)
Frederick

Share this post


Link to post
Share on other sites
Hey kRogue,

maybe this OpenGL.org thread is of interest for you, maybe you've already read it =) :
VBO layout of attributes

There is a lot about optimal cache utilization, vbo usage etc. Unfortunately the end of thread is disturbing to me, one of the participants has the same problem as me, a maximum batch size of 4096 indices - unfortunately the problem remains unresolved. This is a direct contradiction to the best practice of creating large batches, isn't it ?

Share this post


Link to post
Share on other sites
Hi,

On my GeForce 8700M GT, I get 1048576 for both (this value is 2^20)... me thinks that there is a driver bug when it reports 4096(=2^12) for both. One thing I do know is that one should use 16bit indices in the glDrawRangeElements calls, which means you can have up yo 65536 distinct vertices for one glDrawRangeElements call. This I do, I also re-order the indices to prime the vertex cache.

But this all does not address the initial question, what kind of performance gains are there with 1 large VBO for many objects vs one (or several) VBO's for each objects...again, since indices have to be 16bit integers it is inadvisable to bind the one huge VBO and call glVertexAttrib once and rely on funky indices to do the right thing, so for each mesh one has to call glVertexAttrib which, according to nVidia docs is pricier than binding a VBO.




Share this post


Link to post
Share on other sites
Quote:
Original post by kRogue
On my GeForce 8700M GT, I get 1048576 for both (this value is 2^20)... me thinks that there is a driver bug when it reports 4096(=2^12) for both.

Err, I got 150,000 for indices and 2048 for vertices on a 8800GT! Is that realistic?

Quote:
Original post by kRogue
[...] since indices have to be 16bit integers it is inadvisable to bind the one huge VBO and call glVertexAttrib once and rely on funky indices to do the right thing, so for each mesh one has to call glVertexAttrib which, according to nVidia docs is pricier than binding a VBO.

Not necessarily. Using 2MB somewhere between the advised 1MB to 4MB, and dividing that by 65536, you'll get 32 bytes per vertex. E.g. 12 bytes for the position, another 12 bytes for the normal, and 8 bytes for a single texture, and you've already summed up to 32 bytes.

Share this post


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

  • Advertisement
  • Advertisement
  • Popular Tags

  • Similar Content

    • By McGrane
      Hey
      My laptop recently decided to die, so Ive been transferring my project to my work laptop just to get it up to date, and commit it. I was banging my head against the wall all day, as my textures where not displaying in my program- I was getting no errors and no indication of why it was occurring so I have been just trying to figure it out- I know the image loading was working ok, as im using image data elsewhere, I was pretty confident that the code was fine also, as ive never had an issue with displaying textures before, so I thought it might be the drivers on this laptop, (my old one was just using the built in IntelHD, while this laptop has a NVIDIA graphics card) but all seems to be up to date.
      Below are my basic shaders:
      Vertex Shader
      #version 330 core layout(location = 0) in vec3 position; layout(location = 1) in vec3 color; layout(location = 2) in vec3 normal; layout(location = 3) in vec2 texCoord; uniform mat4 Projection; uniform mat4 Model; out vec3 Color; out vec3 Normal; out vec2 TexCoord; void main() { gl_Position = Projection * Model * vec4( position, 1.0 ); Color = color; Normal = normal; TexCoord = vec2( texCoord.x, texCoord.y); } Fragment Shader
      #version 330 core in vec3 Color; in vec3 Normal; in vec2 TexCoord; uniform sampler2D textureData; void main() { vec4 textureColor = texture( textureData, TexCoord ); vec4 finalColor = textureColor * vec4( Color, 1.0f); gl_FragColor = finalColor; } Calling Code
      glActiveTexture(GL_TEXTURE0); glBindTexture(GL_TEXTURE_2D, textureID); glUniform1i(glGetUniformLocation(shaderID, "textureData"), textureID); Now this is the part i dont understand, I worked through my program, until I got to the above 'Calling Code'. This just displays a black texture.. my original issue. Out of desperation, I just tried changing the name in glGetUniformLocation from "textureData" to "textureData_invalid" to see if my error checks would through up something, but in actual fact, it is now displaying the texture as expected. Can anyone fathom a guess as too why this is occurring.. im assuming the random text is just picking up the correct location by c++ witchcraft, but why is the original one not getting picked up correctly and/or not working as expected
      I realize more code is probably needed to see how it all hangs together.. but it seems to come down to this as the issue
    • By QQemka
      Hello. So far i got decently looking 3d scene. I also managed to render a truetype font, on my way to implementing gui (windows, buttons and textboxes). There are several issues i am facing, would love to hear your feedback.
      1) I render text using atlas with VBO containing x/y/u/v of every digit in the atlas (calculated basing on x/y/z/width/height/xoffset/yoffset/xadvance data in binary .fnt format file, screenshot 1). I generated a Comic Sans MS with 32 size and times new roman with size 12 (screenshot 2 and 3). The first issue is the font looks horrible when rescaling. I guess it is because i am using fixed -1 to 1 screen space coords. This is where ortho matrix should be used, right?
      2) Rendering GUI. Situation is similar to above. I guess the widgets should NOT scale when scaling window, am i right? So what am i looking for is saying "this should be always in the middle, 200x200 size no matter the display window xy", and "this should stick to the bottom left corner". Is ortho matrix the cure for all such problems?
      3) The game is 3D but i have to go 2D to render static gui elements over the scene - and i want to do it properly! At the moment i am using matrix 3x3 for 2d transformations and vec3 for all kinds of coordinates. In shaders tho i technically still IS 3D. I have to set all 4 x y z w of the gl_Position while it would be much much more conventient to... just do the maths in 2d space. Can i achieve it somehow?
      4) Text again. I am kind of confused what is the reason of artifacts in Times New Roman font displaying (screenshot 1). I render from left to right, letter after letter. You can clearly see that letters on the right (so the ones rendered after ones on the left are covered by the previous one). I was toying around with blending options but no luck. I do not support kerning at the moment but that's definitely not the cause of error. The display of the small font looks dirty aliased too. I am knd of confused how to interpret the integer data and how should be scaled/adapted to the screen view. Is it just store the data as constant size and again - use ortho matrix?
      Thanks in advance for all your ideas and suggestions!
      https://i.imgur.com/4rd1VC3.png
      https://i.imgur.com/uHrSXfe.png
      https://i.imgur.com/xRTffPn.png
    • By plz717
      Hello, everyone! I hope my problem isn't too 'beginnerish'. I'm doing research on motion synthesis now, trying to implement the Deep Mimic paper (DeepMimic) by BINPENG XUE, in this paper, I need to first retarget character A's motion to another character B to make the reference motion clips for character B, since we don't have character B‘s reference motion. The most important thing is that in the paper, the author copied character A's joint's rotation with respective to joint's local coordinate system (not the parent) to character B. In my personal understanding, the joint's rotation with respective to joint's local coordinate system is something like that in the attached photo, where for the Elbow joint, i need to get the Elbow's rotation in the elbow's local coordinate system (i'm very grateful for you to share your ideas if i have misunderstanding about it 🙂)
      I have searched many materials on the internet about how to extract the local joint's information from FBX, the most relative one i found is the pivot rotation( and geometric transformation, object offset transformation). I'm a beginner in computer graphics, and i'm confused about whether the pivot rotation( or geometric transformation, object offset transformation) is exactly the joint's local rotation i'm seeking? I hope someone that have any ideas can help me, I'd be very grateful for any pointers in the right direction. Thanks in advance! 

    • By nOoNEE
      hello guys , i have some questions  what does glLinkProgram  and  glBindAttribLocation do?  i searched but there wasnt any good resource 
    • By owenjr
      Hi, I'm a Multimedia Engineering student. I am about to finish my dergree and I'm already thinking about what topic to cover in my final college project.
      I'm interested in the procedural animation with c++ and OpenGL of creatures, something like a spider for example. Can someone tell me what are the issues I should investigate to carry it out? I understand that it has some dependence on artificial intelligence but I do not know to what extent. Can someone help me to find information about it? Thank you very much.
       
      Examples: 
      - Procedural multi-legged walking animation
      - Procedural Locomotion of Multi-Legged Characters in Dynamic Environments
    • By Lewa
      So, i'm still on my quest to unterstanding the intricacies of HDR and implementing this into my engine. Currently i'm at the step to implementing tonemapping. I stumbled upon this blogposts:
      http://filmicworlds.com/blog/filmic-tonemapping-operators/
      http://frictionalgames.blogspot.com/2012/09/tech-feature-hdr-lightning.html
      and tried to implement some of those mentioned tonemapping methods into my postprocessing shader.
      The issue is that none of them creates the same results as shown in the blogpost which definitely has to do with the initial range in which the values are stored in the HDR buffer. For simplicity sake i store the values between 0 and 1 in the HDR buffer (ambient light is 0.3, directional light is 0.7)
      This is the tonemapping code:
      vec3 Uncharted2Tonemap(vec3 x) { float A = 0.15; float B = 0.50; float C = 0.10; float D = 0.20; float E = 0.02; float F = 0.30; return ((x*(A*x+C*B)+D*E)/(x*(A*x+B)+D*F))-E/F; } This is without the uncharted tonemapping:
      This is with the uncharted tonemapping:
      Which makes the image a lot darker.
      The shader code looks like this:
      void main() { vec3 color = texture2D(texture_diffuse, vTexcoord).rgb; color = Uncharted2Tonemap(color); //gamma correction (use only if not done in tonemapping code) color = gammaCorrection(color); outputF = vec4(color,1.0f); } Now, from my understanding is that tonemapping should bring the range down from HDR to 0-1.
      But the output of the tonemapping function heavily depends on the initial range of the values in the HDR buffer. (You can't expect to set the sun intensity the first time to 10 and the second time to 1000 and excpect the same result if you feed that into the tonemapper.) So i suppose that this also depends on the exposure which i have to implement?
      To check this i plotted the tonemapping curve:
      You can see that the curve goes only up to around to a value of 0.21 (while being fed a value of 1) and then basically flattens out. (which would explain why the image got darker.)
       
      My guestion is: In what range should the values in the HDR buffer be which then get tonemapped? Do i have to bring them down to a range of 0-1 by multiplying with the exposure?
      For example, if i increase the values of the light by 10 (directional light would be 7 and ambient light 3) then i would need to divide HDR values by 10 in order to get a value range of 0-1 which then could be fed into the tonemapping curve. Is that correct?
    • By nOoNEE
      i am reading this book : link
      in the OpenGL Rendering Pipeline section there is a picture like this: link
      but the question is this i dont really understand why it is necessary to turn pixel data in to fragment and then fragment into pixel could please give me a source or a clear Explanation that why it is necessary ? thank you so mu
       
       
    • By Inbar_xz
      I'm using the OPENGL with eclipse+JOGL.
      My goal is to create movement of the camera and the player.
      I create main class, which create some box in 3D and hold 
      an object of PlayerAxis.
      I create PlayerAxis class which hold the axis of the player.
      If we want to move the camera, then in the main class I call to 
      the func "cameraMove"(from PlayerAxis) and it update the player axis.
      That's work good.
      The problem start if I move the camera on 2 axis, 
      for example if I move with the camera right(that's on the y axis)
      and then down(on the x axis) -
      in some point the move front is not to the front anymore..
      In order to move to the front, I do
      player.playerMoving(0, 0, 1);
      And I learn that in order to keep the front move, 
      I need to convert (0, 0, 1) to the player axis, and then add this.
      I think I dont do the convert right.. 
      I will be glad for help!

      Here is part of my PlayerAxis class:
       
      //player coordinate float x[] = new float[3]; float y[] = new float[3]; float z[] = new float[3]; public PlayerAxis(float move_step, float angle_move) { x[0] = 1; y[1] = 1; z[2] = -1; step = move_step; angle = angle_move; setTransMatrix(); } public void cameraMoving(float angle_step, String axis) { float[] new_x = x; float[] new_y = y; float[] new_z = z; float alfa = angle_step * angle; switch(axis) { case "x": new_z = addVectors(multScalar(z, COS(alfa)), multScalar(y, SIN(alfa))); new_y = subVectors(multScalar(y, COS(alfa)), multScalar(z, SIN(alfa))); break; case "y": new_x = addVectors(multScalar(x, COS(alfa)), multScalar(z, SIN(alfa))); new_z = subVectors(multScalar(z, COS(alfa)), multScalar(x, SIN(alfa))); break; case "z": new_x = addVectors(multScalar(x, COS(alfa)), multScalar(y, SIN(alfa))); new_y = subVectors(multScalar(y, COS(alfa)), multScalar(x, SIN(alfa))); } x = new_x; y = new_y; z = new_z; normalization(); } public void playerMoving(float x_move, float y_move, float z_move) { float[] move = new float[3]; move[0] = x_move; move[1] = y_move; move[2] = z_move; setTransMatrix(); float[] trans_move = transVector(move); position[0] = position[0] + step*trans_move[0]; position[1] = position[1] + step*trans_move[1]; position[2] = position[2] + step*trans_move[2]; } public void setTransMatrix() { for (int i = 0; i < 3; i++) { coordiTrans[0][i] = x[i]; coordiTrans[1][i] = y[i]; coordiTrans[2][i] = z[i]; } } public float[] transVector(float[] v) { return multiplyMatrixInVector(coordiTrans, v); }  
      and in the main class i have this:
       
      public void keyPressed(KeyEvent e) { if (e.getKeyCode()== KeyEvent.VK_ESCAPE) { System.exit(0); //player move } else if (e.getKeyCode()== KeyEvent.VK_W) { //front //moveAmount[2] += -0.1f; player.playerMoving(0, 0, 1); } else if (e.getKeyCode()== KeyEvent.VK_S) { //back //moveAmount[2] += 0.1f; player.playerMoving(0, 0, -1); } else if (e.getKeyCode()== KeyEvent.VK_A) { //left //moveAmount[0] += -0.1f; player.playerMoving(-1, 0, 0); } else if (e.getKeyCode()== KeyEvent.VK_D) { //right //moveAmount[0] += 0.1f; player.playerMoving(1, 0, 0); } else if (e.getKeyCode()== KeyEvent.VK_E) { //moveAmount[0] += 0.1f; player.playerMoving(0, 1, 0); } else if (e.getKeyCode()== KeyEvent.VK_Q) { //moveAmount[0] += 0.1f; player.playerMoving(0, -1, 0); //camera move } else if (e.getKeyCode()== KeyEvent.VK_I) { //up player.cameraMoving(1, "x"); } else if (e.getKeyCode()== KeyEvent.VK_K) { //down player.cameraMoving(-1, "x"); } else if (e.getKeyCode()== KeyEvent.VK_L) { //right player.cameraMoving(-1, "y"); } else if (e.getKeyCode()== KeyEvent.VK_J) { //left player.cameraMoving(1, "y"); } else if (e.getKeyCode()== KeyEvent.VK_O) { //right round player.cameraMoving(-1, "z"); } else if (e.getKeyCode()== KeyEvent.VK_U) { //left round player.cameraMoving(1, "z"); } }  
      finallt found it.... i confused with the transformation matrix row and col. thanks anyway!
    • By Lewa
      So, i'm currently trying to implement an SSAO shader from THIS tutorial and i'm running into a few issues here.
      Now, this SSAO method requires view space positions and normals. I'm storing the normals in my deferred renderer in world-space so i had to do a conversion and reconstruct the position from the depth buffer.
      And something there goes horribly wrong (which has probably to do with worldspace to viewspace transformations).
      (here is the full shader source code if someone wants to take a look at it)
      Now, i suspect that the normals are the culprit.
      vec3 normal = ((uNormalViewMatrix*vec4(normalize(texture2D(sNormals, vTexcoord).rgb),1.0)).xyz); "sNormals" is a 2D texture which stores the normals in world space in a RGB FP16 buffer.
      Now i can't use the camera viewspace matrix to transform the normals into viewspace as the cameras position isn't set at (0,0,0), thus skewing the result.
      So what i did is to create a new viewmatrix specifically for this normal without the position at vec3(0,0,0);
      //"camera" is the camera which was used for rendering the normal buffer renderer.setUniform4m(ressources->shaderSSAO->getUniform("uNormalViewMatrix"), glmExt::createViewMatrix(glm::vec3(0,0,0),camera.getForward(),camera.getUp())//parameters are (position,forwardVector,upVector) ); Though i have the feeling this is the wrong approach. Is this right or is there a better/correct way of transforming a world space normal into viewspace?
    • By HawkDeath
      Hi,
      I'm trying mix two textures using own shader system, but I have a problem (I think) with uniforms.
      Code: https://github.com/HawkDeath/shader/tree/test
      To debug I use RenderDocs, but I did not receive good results. In the first attachment is my result, in the second attachment is what should be.
      PS. I base on this tutorial https://learnopengl.com/Getting-started/Textures.


  • Advertisement
  • Popular Now

  • Forum Statistics

    • Total Topics
      631399
    • Total Posts
      2999854
×

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!