Jump to content
  • Advertisement
Sign in to follow this  
Heodox

OpenGL Skinning in software

This topic is 3639 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

I'm trying to implement model skinning in software opengl (not using shaders). The animation works, the problem is the speed, first it was done directly by calling glbegin/glend , and using glVertex... This was SLOW (~10 fps), now i implemented it with glDrawArrays and it is still prety slow (~25 fps). the skin data is stored along with vertex data in standard shader skinning format (one bone_index vec4, and one weight_index vec4) I'm wondering what would be the best/fastes way of doing this?

Share this post


Link to post
Share on other sites
Advertisement
Get a profiler and figure out where the bottleneck is. It's possible that your GPU is faster at transforming verts than your CPU is.

Why are you trying to do software skinning? How long was it taking the GPU to skin the verts? Was the bottleneck really the transform?

Share this post


Link to post
Share on other sites
Quote:
Original post by RDragon1
Get a profiler and figure out where the bottleneck is. It's possible that your GPU is faster at transforming verts than your CPU is.




i think the biggest bottle neck is that i calculate every vertex like this on CPU:

v += mat[ind1] * original_v * weight1
v += mat[ind2] * original_v * weight2
v += mat[ind3] * original_v * weight3
v += mat[ind4] * original_v * weight4

but if i remove this (the model is not moving) i still get prety bad results (200-400-600 fps, very jumpy ), also i get very jumpy results for display lists 200-400-800 fps.

Quote:
Original post by RDragon1
Why are you trying to do software skinning? How long was it taking the GPU to skin the verts? Was the bottleneck really the transform?


I'm trying to add as much compatibility to the engine as i can (this would mainly be character animation without shaders), the skinning on gpu is not implemented yet but i know it will be fast (did it on last project).

I belive the bottlenect mostly to be transform, but i also now think there is something else slowing it down.


also what profiler would you recomend? (i'm using visual studio 2008 if that is of any importance)

Share this post


Link to post
Share on other sites
You would need to optimize it with some intrinsics or inline assembly SSE.

Share this post


Link to post
Share on other sites
Quote:
Original post by V-man
You would need to optimize it with some intrinsics or inline assembly SSE.


It's unlikely to give you a huge boost to frame rates though, chances are the bottlenecks are 1) the amount of memory that needs to be read/written 2) the time it takes to upload the vertex array to the graphics card. SSE might give you slightly better performance, however simply using openMP and a #pragma omp parallel for would be a better idea. Better still would be to utilise data compression for the vtx data to minimise the amount of data you need to read.

Quote:

I belive the bottlenect mostly to be transform, but i also now think there is something else slowing it down.


Software skinning shouldn't be that slow, i can do around 1 million verts at playable frame rates (though it is admittedly heavily optimised and multi-threaded). I suspect you have something nasty going on such as a floating point exception or division by zero. Try turning on all exceptions in your debugger and see if anything gets triggered. Alternatively vtune would probably be able to identify the problem (probably the best tool for this job - assuming you have an intel processor).

Share this post


Link to post
Share on other sites
Quick *profiling* could be also disabling all drawing and check CPU only fps. This way you can quickly confirm if it is CPU or transfer/GPU bottleneck.

Share this post


Link to post
Share on other sites
Quote:
Original post by MaR_dev
Quick *profiling* could be also disabling all drawing and check CPU only fps. This way you can quickly confirm if it is CPU or transfer/GPU bottleneck.


The fps doesn't change if i turn off the the draw calls, so it is purely cpu bottleneck.

I've managed to do a couple optimizations and now debug version has 45 fps and release version 350 fps.

I see one more big optimization that i can do is to compress vertex data (as suggested by RobTheBloke), but else from that i can't see anything more i can do (profiler showed me some stupid stuff i did, and now that is removed)

Share this post


Link to post
Share on other sites
Quote:
Original post by Heodox
The fps doesn't change if i turn off the the draw calls, so it is purely cpu bottleneck.


Disable all rendering (and buffer swapping), then you'll have a true indication of the performance. Don't forget that any SwapBuffers call takes a fairly long time to complete..... probably longer than your skinning computation.

I'd also try experimenting with openMP to do something like:


#pragma omp parallel for
for(int i=0;i<num_verts;++i)
{
v = 0;
v += mat[ind1] * original_v * weight1
v += mat[ind2] * original_v * weight2
v += mat[ind3] * original_v * weight3
v += mat[ind4] * original_v * weight4
}


For small data sets, you should see an improvement (i.e. if the skin/vtx data fits nicely in CPU cache) - assuming of course you have a dual or quad core CPU... (make sure you enable openMP in your project settings...)

If you mesh data is large, you won't see any difference and compression is your only option.

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 mmmax3d
      Hi everyone,
      I would need some assistance from anyone who has a similar experience
      or a nice idea!
      I have created a skybox (as cube) and now I need to add a floor/ground.
      The skybox is created from cubemap and initially it was infinite.
      Now it is finite with a specific size. The floor is a quad in the middle
      of the skybox, like a horizon.
      I have two problems:
      When moving the skybox upwards or downwards, I need to
      sample from points even above the horizon while sampling
      from the botton at the same time.  I am trying to create a seamless blending of the texture
      at the points of the horizon, when the quad is connected
      to the skybox. However, I get skew effects. Does anybody has done sth similar?
      Is there any good practice?
      Thanks everyone!
    • By mmmax3d
      Hi everyone,
      I would need some assistance from anyone who has a similar experience
      or a nice idea!
      I have created a skybox (as cube) and now I need to add a floor/ground.
      The skybox is created from cubemap and initially it was infinite.
      Now it is finite with a specific size. The floor is a quad in the middle
      of the skybox, like a horizon.
      I have two problems:
      When moving the skybox upwards or downwards, I need to
      sample from points even above the horizon while sampling
      from the botton at the same time.  I am trying to create a seamless blending of the texture
      at the points of the horizon, when the quad is connected
      to the skybox. However, I get skew effects. Does anybody has done sth similar?
      Is there any good practice?
      Thanks everyone!
    • By iArtist93
      I'm trying to implement PBR into my simple OpenGL renderer and trying to use multiple lighting passes, I'm using one pass per light for rendering as follow:
      1- First pass = depth
      2- Second pass = ambient
      3- [3 .. n] for all the lights in the scene.
      I'm using the blending function glBlendFunc(GL_ONE, GL_ONE) for passes [3..n], and i'm doing a Gamma Correction at the end of each fragment shader.
      But i still have a problem with the output image it just looks noisy specially when i'm using texture maps.
      Is there anything wrong with those steps or is there any improvement to this process?
    • By babaliaris
      Hello Everyone!
      I'm learning openGL, and currently i'm making a simple 2D game engine to test what I've learn so far.  In order to not say to much, i made a video in which i'm showing you the behavior of the rendering.
      Video: 
       
      What i was expecting to happen, was the player moving around. When i render only the player, he moves as i would expect. When i add a second Sprite object, instead of the Player, this new sprite object is moving and finally if i add a third Sprite object the third one is moving. And the weird think is that i'm transforming the Vertices of the Player so why the transformation is being applied somewhere else?
       
      Take a look at my code:
      Sprite Class
      (You mostly need to see the Constructor, the Render Method and the Move Method)
      #include "Brain.h" #include <glm/gtc/matrix_transform.hpp> #include <vector> struct Sprite::Implementation { //Position. struct pos pos; //Tag. std::string tag; //Texture. Texture *texture; //Model matrix. glm::mat4 model; //Vertex Array Object. VertexArray *vao; //Vertex Buffer Object. VertexBuffer *vbo; //Layout. VertexBufferLayout *layout; //Index Buffer Object. IndexBuffer *ibo; //Shader. Shader *program; //Brains. std::vector<Brain *> brains; //Deconstructor. ~Implementation(); }; Sprite::Sprite(std::string image_path, std::string tag, float x, float y) { //Create Pointer To Implementaion. m_Impl = new Implementation(); //Set the Position of the Sprite object. m_Impl->pos.x = x; m_Impl->pos.y = y; //Set the tag. m_Impl->tag = tag; //Create The Texture. m_Impl->texture = new Texture(image_path); //Initialize the model Matrix. m_Impl->model = glm::mat4(1.0f); //Get the Width and the Height of the Texture. int width = m_Impl->texture->GetWidth(); int height = m_Impl->texture->GetHeight(); //Create the Verticies. float verticies[] = { //Positions //Texture Coordinates. x, y, 0.0f, 0.0f, x + width, y, 1.0f, 0.0f, x + width, y + height, 1.0f, 1.0f, x, y + height, 0.0f, 1.0f }; //Create the Indicies. unsigned int indicies[] = { 0, 1, 2, 2, 3, 0 }; //Create Vertex Array. m_Impl->vao = new VertexArray(); //Create the Vertex Buffer. m_Impl->vbo = new VertexBuffer((void *)verticies, sizeof(verticies)); //Create The Layout. m_Impl->layout = new VertexBufferLayout(); m_Impl->layout->PushFloat(2); m_Impl->layout->PushFloat(2); m_Impl->vao->AddBuffer(m_Impl->vbo, m_Impl->layout); //Create the Index Buffer. m_Impl->ibo = new IndexBuffer(indicies, 6); //Create the new shader. m_Impl->program = new Shader("Shaders/SpriteShader.shader"); } //Render. void Sprite::Render(Window * window) { //Create the projection Matrix based on the current window width and height. glm::mat4 proj = glm::ortho(0.0f, (float)window->GetWidth(), 0.0f, (float)window->GetHeight(), -1.0f, 1.0f); //Set the MVP Uniform. m_Impl->program->setUniformMat4f("u_MVP", proj * m_Impl->model); //Run All The Brains (Scripts) of this game object (sprite). for (unsigned int i = 0; i < m_Impl->brains.size(); i++) { //Get Current Brain. Brain *brain = m_Impl->brains[i]; //Call the start function only once! if (brain->GetStart()) { brain->SetStart(false); brain->Start(); } //Call the update function every frame. brain->Update(); } //Render. window->GetRenderer()->Draw(m_Impl->vao, m_Impl->ibo, m_Impl->texture, m_Impl->program); } void Sprite::Move(float speed, bool left, bool right, bool up, bool down) { if (left) { m_Impl->pos.x -= speed; m_Impl->model = glm::translate(m_Impl->model, glm::vec3(-speed, 0, 0)); } if (right) { m_Impl->pos.x += speed; m_Impl->model = glm::translate(m_Impl->model, glm::vec3(speed, 0, 0)); } if (up) { m_Impl->pos.y += speed; m_Impl->model = glm::translate(m_Impl->model, glm::vec3(0, speed, 0)); } if (down) { m_Impl->pos.y -= speed; m_Impl->model = glm::translate(m_Impl->model, glm::vec3(0, -speed, 0)); } } void Sprite::AddBrain(Brain * brain) { //Push back the brain object. m_Impl->brains.push_back(brain); } pos *Sprite::GetPos() { return &m_Impl->pos; } std::string Sprite::GetTag() { return m_Impl->tag; } int Sprite::GetWidth() { return m_Impl->texture->GetWidth(); } int Sprite::GetHeight() { return m_Impl->texture->GetHeight(); } Sprite::~Sprite() { delete m_Impl; } //Implementation Deconstructor. Sprite::Implementation::~Implementation() { delete texture; delete vao; delete vbo; delete layout; delete ibo; delete program; }  
      Renderer Class
      #include "Renderer.h" #include "Error.h" Renderer::Renderer() { } Renderer::~Renderer() { } void Renderer::Draw(VertexArray * vao, IndexBuffer * ibo, Texture *texture, Shader * program) { vao->Bind(); ibo->Bind(); program->Bind(); if (texture != NULL) texture->Bind(); GLCall(glDrawElements(GL_TRIANGLES, ibo->GetCount(), GL_UNSIGNED_INT, NULL)); } void Renderer::Clear(float r, float g, float b) { GLCall(glClearColor(r, g, b, 1.0)); GLCall(glClear(GL_COLOR_BUFFER_BIT)); } void Renderer::Update(GLFWwindow *window) { /* Swap front and back buffers */ glfwSwapBuffers(window); /* Poll for and process events */ glfwPollEvents(); }  
      Shader Code
      #shader vertex #version 330 core layout(location = 0) in vec4 aPos; layout(location = 1) in vec2 aTexCoord; out vec2 t_TexCoord; uniform mat4 u_MVP; void main() { gl_Position = u_MVP * aPos; t_TexCoord = aTexCoord; } #shader fragment #version 330 core out vec4 aColor; in vec2 t_TexCoord; uniform sampler2D u_Texture; void main() { aColor = texture(u_Texture, t_TexCoord); } Also i'm pretty sure that every time i'm hitting the up, down, left and right arrows on the keyboard, i'm changing the model Matrix of the Player and not the others.
       
      Window Class:
      #include "Window.h" #include <GL/glew.h> #include <GLFW/glfw3.h> #include "Error.h" #include "Renderer.h" #include "Scene.h" #include "Input.h" //Global Variables. int screen_width, screen_height; //On Window Resize. void OnWindowResize(GLFWwindow *window, int width, int height); //Implementation Structure. struct Window::Implementation { //GLFW Window. GLFWwindow *GLFW_window; //Renderer. Renderer *renderer; //Delta Time. double delta_time; //Frames Per Second. int fps; //Scene. Scene *scnene; //Input. Input *input; //Deconstructor. ~Implementation(); }; //Window Constructor. Window::Window(std::string title, int width, int height) { //Initializing width and height. screen_width = width; screen_height = height; //Create Pointer To Implementation. m_Impl = new Implementation(); //Try initializing GLFW. if (!glfwInit()) { std::cout << "GLFW could not be initialized!" << std::endl; std::cout << "Press ENTER to exit..." << std::endl; std::cin.get(); exit(-1); } //Setting up OpenGL Version 3.3 Core Profile. glfwWindowHint(GLFW_CONTEXT_VERSION_MAJOR, 3); glfwWindowHint(GLFW_CONTEXT_VERSION_MINOR, 3); glfwWindowHint(GLFW_OPENGL_PROFILE, GLFW_OPENGL_CORE_PROFILE); /* Create a windowed mode window and its OpenGL context */ m_Impl->GLFW_window = glfwCreateWindow(width, height, title.c_str(), NULL, NULL); if (!m_Impl->GLFW_window) { std::cout << "GLFW could not create a window!" << std::endl; std::cout << "Press ENTER to exit..." << std::endl; std::cin.get(); glfwTerminate(); exit(-1); } /* Make the window's context current */ glfwMakeContextCurrent(m_Impl->GLFW_window); //Initialize GLEW. if(glewInit() != GLEW_OK) { std::cout << "GLEW could not be initialized!" << std::endl; std::cout << "Press ENTER to exit..." << std::endl; std::cin.get(); glfwTerminate(); exit(-1); } //Enabling Blending. GLCall(glEnable(GL_BLEND)); GLCall(glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA)); //Setting the ViewPort. GLCall(glViewport(0, 0, width, height)); //**********Initializing Implementation**********// m_Impl->renderer = new Renderer(); m_Impl->delta_time = 0.0; m_Impl->fps = 0; m_Impl->input = new Input(this); //**********Initializing Implementation**********// //Set Frame Buffer Size Callback. glfwSetFramebufferSizeCallback(m_Impl->GLFW_window, OnWindowResize); } //Window Deconstructor. Window::~Window() { delete m_Impl; } //Window Main Loop. void Window::MainLoop() { //Time Variables. double start_time = 0, end_time = 0, old_time = 0, total_time = 0; //Frames Counter. int frames = 0; /* Loop until the user closes the window */ while (!glfwWindowShouldClose(m_Impl->GLFW_window)) { old_time = start_time; //Total time of previous frame. start_time = glfwGetTime(); //Current frame start time. //Calculate the Delta Time. m_Impl->delta_time = start_time - old_time; //Get Frames Per Second. if (total_time >= 1) { m_Impl->fps = frames; total_time = 0; frames = 0; } //Clearing The Screen. m_Impl->renderer->Clear(0, 0, 0); //Render The Scene. if (m_Impl->scnene != NULL) m_Impl->scnene->Render(this); //Updating the Screen. m_Impl->renderer->Update(m_Impl->GLFW_window); //Increasing frames counter. frames++; //End Time. end_time = glfwGetTime(); //Total time after the frame completed. total_time += end_time - start_time; } //Terminate GLFW. glfwTerminate(); } //Load Scene. void Window::LoadScene(Scene * scene) { //Set the scene. m_Impl->scnene = scene; } //Get Delta Time. double Window::GetDeltaTime() { return m_Impl->delta_time; } //Get FPS. int Window::GetFPS() { return m_Impl->fps; } //Get Width. int Window::GetWidth() { return screen_width; } //Get Height. int Window::GetHeight() { return screen_height; } //Get Input. Input * Window::GetInput() { return m_Impl->input; } Renderer * Window::GetRenderer() { return m_Impl->renderer; } GLFWwindow * Window::GetGLFWindow() { return m_Impl->GLFW_window; } //Implementation Deconstructor. Window::Implementation::~Implementation() { delete renderer; delete input; } //OnWindowResize void OnWindowResize(GLFWwindow *window, int width, int height) { screen_width = width; screen_height = height; //Updating the ViewPort. GLCall(glViewport(0, 0, width, height)); }  
      Brain Class
      #include "Brain.h" #include "Sprite.h" #include "Window.h" struct Brain::Implementation { //Just A Flag. bool started; //Window Pointer. Window *window; //Sprite Pointer. Sprite *sprite; }; Brain::Brain(Window *window, Sprite *sprite) { //Create Pointer To Implementation. m_Impl = new Implementation(); //Initialize Implementation. m_Impl->started = true; m_Impl->window = window; m_Impl->sprite = sprite; } Brain::~Brain() { //Delete Pointer To Implementation. delete m_Impl; } void Brain::Start() { } void Brain::Update() { } Window * Brain::GetWindow() { return m_Impl->window; } Sprite * Brain::GetSprite() { return m_Impl->sprite; } bool Brain::GetStart() { return m_Impl->started; } void Brain::SetStart(bool value) { m_Impl->started = value; } Script Class (Its a Brain Subclass!!!)
      #include "Script.h" Script::Script(Window *window, Sprite *sprite) : Brain(window, sprite) { } Script::~Script() { } void Script::Start() { std::cout << "Game Started!" << std::endl; } void Script::Update() { Input *input = this->GetWindow()->GetInput(); Sprite *sp = this->GetSprite(); //Move this sprite. this->GetSprite()->Move(200 * this->GetWindow()->GetDeltaTime(), input->GetKeyDown("left"), input->GetKeyDown("right"), input->GetKeyDown("up"), input->GetKeyDown("down")); std::cout << sp->GetTag().c_str() << ".x = " << sp->GetPos()->x << ", " << sp->GetTag().c_str() << ".y = " << sp->GetPos()->y << std::endl; }  
      Main:
      #include "SpaceShooterEngine.h" #include "Script.h" int main() { Window w("title", 600,600); Scene *scene = new Scene(); Sprite *player = new Sprite("Resources/Images/player.png", "Player", 100,100); Sprite *other = new Sprite("Resources/Images/cherno.png", "Other", 400, 100); Sprite *other2 = new Sprite("Resources/Images/cherno.png", "Other", 300, 400); Brain *brain = new Script(&w, player); player->AddBrain(brain); scene->AddSprite(player); scene->AddSprite(other); scene->AddSprite(other2); w.LoadScene(scene); w.MainLoop(); return 0; }  
       
      I literally can't find what is wrong. If you need more code, ask me to post it. I will also attach all the source files.
      Brain.cpp
      Error.cpp
      IndexBuffer.cpp
      Input.cpp
      Renderer.cpp
      Scene.cpp
      Shader.cpp
      Sprite.cpp
      Texture.cpp
      VertexArray.cpp
      VertexBuffer.cpp
      VertexBufferLayout.cpp
      Window.cpp
      Brain.h
      Error.h
      IndexBuffer.h
      Input.h
      Renderer.h
      Scene.h
      Shader.h
      SpaceShooterEngine.h
      Sprite.h
      Texture.h
      VertexArray.h
      VertexBuffer.h
      VertexBufferLayout.h
      Window.h
    • By Cristian Decu
      Hello fellow programmers,
      For a couple of days now i've decided to build my own planet renderer just to see how floating point precision issues
      can be tackled. As you probably imagine, i've quickly faced FPP issues when trying to render absurdly large planets.
       
      I have used the classical quadtree LOD approach;
      I've generated my grids with 33 vertices, (x: -1 to 1, y: -1 to 1, z = 0).
      Each grid is managed by a TerrainNode class that, depending on the side it represents (top, bottom, left right, front, back),
      creates a special rotation-translation matrix that moves and rotates the grid away from the origin so that when i finally
      normalize all the vertices on my vertex shader i can get a perfect sphere.
      T = glm::translate(glm::dmat4(1.0), glm::dvec3(0.0, 0.0, 1.0)); R = glm::rotate(glm::dmat4(1.0), glm::radians(180.0), glm::dvec3(1.0, 0.0, 0.0)); sides[0] = new TerrainNode(1.0, radius, T * R, glm::dvec2(0.0, 0.0), new TerrainTile(1.0, SIDE_FRONT)); T = glm::translate(glm::dmat4(1.0), glm::dvec3(0.0, 0.0, -1.0)); R = glm::rotate(glm::dmat4(1.0), glm::radians(0.0), glm::dvec3(1.0, 0.0, 0.0)); sides[1] = new TerrainNode(1.0, radius, R * T, glm::dvec2(0.0, 0.0), new TerrainTile(1.0, SIDE_BACK)); // So on and so forth for the rest of the sides As you can see, for the front side grid, i rotate it 180 degrees to make it face the camera and push it towards the eye;
      the back side is handled almost the same way only that i don't need to rotate it but simply push it away from the eye.
      The same technique is applied for the rest of the faces (obviously, with the proper rotations / translations).
      The matrix that result from the multiplication of R and T (in that particular order) is send to my vertex shader as `r_Grid'.
      // spherify vec3 V = normalize((r_Grid * vec4(r_Vertex, 1.0)).xyz); gl_Position = r_ModelViewProjection * vec4(V, 1.0); The `r_ModelViewProjection' matrix is generated on the CPU in this manner.
      // No the most efficient way, but it works. glm::dmat4 Camera::getMatrix() { // Create the view matrix // Roll, Yaw and Pitch are all quaternions. glm::dmat4 View = glm::toMat4(Roll) * glm::toMat4(Pitch) * glm::toMat4(Yaw); // The model matrix is generated by translating in the oposite direction of the camera. glm::dmat4 Model = glm::translate(glm::dmat4(1.0), -Position); // Projection = glm::perspective(fovY, aspect, zNear, zFar); // zNear = 0.1, zFar = 1.0995116e12 return Projection * View * Model; } I managed to get rid of z-fighting by using a technique called Logarithmic Depth Buffer described in this article; it works amazingly well, no z-fighting at all, at least not visible.
      Each frame i'm rendering each node by sending the generated matrices this way.
      // set the r_ModelViewProjection uniform // Sneak in the mRadiusMatrix which is a matrix that contains the radius of my planet. Shader::setUniform(0, Camera::getInstance()->getMatrix() * mRadiusMatrix); // set the r_Grid matrix uniform i created earlier. Shader::setUniform(1, r_Grid); grid->render(); My planet's radius is around 6400000.0 units, absurdly large, but that's what i really want to achieve;
      Everything works well, the node's split and merge as you'd expect, however whenever i get close to the surface
      of the planet the rounding errors start to kick in giving me that lovely stairs effect.
      I've read that if i could render each grid relative to the camera i could get better precision on the surface, effectively
      getting rid of those rounding errors.
       
      My question is how can i achieve this relative to camera rendering in my scenario here?
      I know that i have to do most of the work on the CPU with double, and that's exactly what i'm doing.
      I only use double on the CPU side where i also do most of the matrix multiplications.
      As you can see from my vertex shader i only do the usual r_ModelViewProjection * (some vertex coords).
       
      Thank you for your suggestions!
       
  • 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!