繪制多個重復的模型時，使用Instanced方式繪制可以大大加快顯然速度。

繪制效果如下圖所示：

1、Vertex Shader中定義如下：

attribute vec3 pos;
attribute vec2 coord;
attribute vec3 normal;
attribute vec3 offset;

uniform mat4 M;
uniform mat4 V;
uniform mat4 P;
uniform mat4 NM;

varying vec2 M_coord;
varying vec3 M_normal;
varying vec4 M_WordPos;

void main()
{
        M_coord = coord;
        M_WordPos = M * vec4(pos, 1.0) + vec4(offset, 1.0);
        M_normal = mat3(NM) * normal;
        gl_Position = P * V * M_WordPos;
}

offset為每次繪制的偏移量。

2、CPU端實現

float nIndexOffset[] = {
        -100.0f, 0.0f, 0.0f,
        0.0f, 0.0f, 0.0f,
        100.0f, 0.0f, 0.0f
    };

createBufferObject為封裝創建OpenGL Buffer的函數

定義如下：

GLuint OpenGLOperate::createBufferObject(GLenum nBufferType, GLsizeiptr size,                                             const GLvoid *data, GLenum usage)
{
    GLuint BufferObjIndex = 0;
    m_OpenGLCore->glGenBuffers(1, &BufferObjIndex);
    m_OpenGLCore->glBindBuffer(nBufferType, BufferObjIndex);
    m_OpenGLCore->glBufferData(nBufferType, size, data, usage);
    m_OpenGLCore->glBindBuffer(nBufferType, 0);
    return BufferObjIndex;
}

創建Offset的FBO

m_OffsetIndex = m_OpenGLOperate->createBufferObject(GL_ARRAY_BUFFER,                                 sizeof(float) * 9,                                 (void*)nIndexOffset, GL_STATIC_DRAW);

為Offset賦值

OpenGLCore->glBindBuffer(GL_ARRAY_BUFFER, m_OffsetIndex);
OpenGLCore->glEnableVertexAttribArray(m_OffsetLocation);
OpenGLCore->glVertexAttribPointer(m_OffsetLocation, 3, GL_FLOAT, GL_FALSE, sizeof(float) * 3, (void*)0);
OpenGLCore->glBindBuffer(GL_ARRAY_BUFFER, 0);
OpenGLCore->glVertexAttribDivisor(m_OffsetLocation, 1);

調用繪制指令：

OpenGLCore->glDrawElementsInstanced(GL_TRIANGLES, m_VertexCount,                                     GL_UNSIGNED_INT, 0, 3);

本文出自 “不會飛的紙飛機” 博客，請務必保留此出處http://douzhq.blog.51cto.com/12552184/1930997

基於Qt的OpenGL可編程管線學習（3）- 使用Instanced方式繪制