后续:异步屏幕外查询性能
Follow up: Asynchronous off-screen query performance
我最近问了这个问题:
如何执行异步外部查询?
我听到的,但尚未确认,窗户上的渲染比渲染到Framebuffer更昂贵。首先,谁能对此发表评论?我可以绘制多个场景以比窗口更快地绘制框架吗?还有其他选项,例如pbuffers还是PBO?
我已经开始玩Framebuffers,但是我无法让查询工作。这是我到目前为止设置的一些伪造代码:
glfwWindowHint(GLFW_VISIBLE, GL_FALSE);
window = glfwCreateWindow(1, 1, "OpenGL", NULL, NULL);
glfwMakeContextCurrent(window);
glGenFramebuffers(1, &fbo);
glBindFramebuffer(GL_DRAW_FRAMEBUFFER, fbo);
glGenRenderbuffers(1, &rbo);
glBindRenderbuffer(GL_RENDERBUFFER, rbo);
glRenderbufferStorage(GL_RENDERBUFFER, GL_DEPTH_COMPONENT24, size, size);
glFramebufferRenderbuffer(GL_DRAW_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_RENDERBUFFER, rbo);
glEnable(GL_DEPTH_TEST);
glGenQueries(numberOfQueries, queries);
for (scene in scenesToRender)
{
glClear(GL_DEPTH_BUFFER_BIT);
glDepthFunc(GL_LESS);
drawShadingObjects(scene);
glBeginQuery(GL_SAMPLES_PASSED, queries[index]);
glDepthFunc(GL_LEQUAL);
drawShadedObject(scene);
glEndQuery(GL_SAMPLES_PASSED);
}
collectQueryResults();
deleteBuffers();
到目前为止一切都运行,但是所有查询都返回" 0"。绘制到框架缓冲器时与绘制窗口缓冲区不同的绘制时是否有一些查询?
再次,我的两个问题是:
- 我可以绘制多个场景以比窗口更快地绘制框架吗?还有其他选项,例如pbuffers还是PBO?
- 在绘制到框架缓冲器时与绘制窗口缓冲区时不同的查询是否有一些?
尝试这样的东西:
#include <GL/glew.h>
#include <GLFW/glfw3.h>
#include <iostream>
#include <vector>
using namespace std;
const unsigned int sz = 1024;
void drawScene( unsigned int multiplier )
{
glViewport( 0, 0, sz, sz );
glClear( GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT );
glMatrixMode(GL_PROJECTION);
glLoadIdentity();
glMatrixMode(GL_MODELVIEW);
glLoadIdentity();
glRotatef( (float)glfwGetTime() * 50.f * multiplier, 0.f, 0.f, 1.f);
glBegin(GL_TRIANGLES);
glColor3f(1.f, 0.f, 0.f);
glVertex3f(-0.6f, -0.4f, 0.f);
glColor3f(0.f, 1.f, 0.f);
glVertex3f(0.6f, -0.4f, 0.f);
glColor3f(0.f, 0.f, 1.f);
glVertex3f(0.f, 0.6f, 0.f);
glEnd();
}
bool available( const vector< GLuint >& queries )
{
for( size_t i = 0; i < queries.size(); ++i )
{
GLuint available = 0;
glGetQueryObjectuiv( queries[i], GL_QUERY_RESULT_AVAILABLE, &available );
if( GL_FALSE == available )
return false;
}
return true;
}
int main()
{
glfwInit();
GLFWwindow* window = glfwCreateWindow( 400, 400, "Simple example", NULL, NULL );
glfwMakeContextCurrent( window );
glewInit();
if( !glewIsSupported( "GL_VERSION_2_1" ) )
return -1;
if( !glewIsSupported( "GL_EXT_framebuffer_object" ) )
return -1;
GLuint fbo = 0;
glGenFramebuffersEXT( 1, &fbo );
glBindFramebufferEXT( GL_DRAW_FRAMEBUFFER_EXT, fbo );
GLuint rbo0 = 0;
glGenRenderbuffersEXT( 1, &rbo0 );
glBindRenderbufferEXT( GL_RENDERBUFFER_EXT, rbo0 );
glRenderbufferStorageEXT( GL_RENDERBUFFER_EXT, GL_RGBA, sz, sz );
glFramebufferRenderbufferEXT( GL_DRAW_FRAMEBUFFER_EXT, GL_COLOR_ATTACHMENT0_EXT, GL_RENDERBUFFER_EXT, rbo0 );
GLuint rbo1 = 0;
glGenRenderbuffersEXT( 1, &rbo1 );
glBindRenderbufferEXT( GL_RENDERBUFFER_EXT, rbo1 );
glRenderbufferStorageEXT( GL_RENDERBUFFER_EXT, GL_DEPTH_COMPONENT24, sz, sz );
glFramebufferRenderbufferEXT( GL_DRAW_FRAMEBUFFER_EXT, GL_DEPTH_ATTACHMENT_EXT, GL_RENDERBUFFER_EXT, rbo1 );
GLenum status = glCheckFramebufferStatusEXT( GL_FRAMEBUFFER_EXT );
if( status != GL_FRAMEBUFFER_COMPLETE_EXT )
return -1;
glBindFramebufferEXT( GL_FRAMEBUFFER_EXT, 0 );
vector< GLuint > queries( 10 );
glGenQueries( queries.size(), &queries[0] );
glBindFramebufferEXT( GL_FRAMEBUFFER_EXT, fbo );
for( size_t i = 0; i < queries.size(); ++i )
{
glBeginQuery( GL_SAMPLES_PASSED, queries[i] );
drawScene( i + 1 );
glEndQuery( GL_SAMPLES_PASSED );
}
glBindFramebufferEXT( GL_FRAMEBUFFER_EXT, 0 );
// wait for queries to become available
unsigned int cnt = 0;
while( !available( queries ) )
{
cnt++;
}
// all queries available, display query results
cout << "cnt: " << cnt << endl;
for( size_t i = 0; i < queries.size(); ++i )
{
GLuint samples = 0;
glGetQueryObjectuiv( queries[i], GL_QUERY_RESULT, &samples );
cout << i << ": " << samples << endl;
}
cout << endl;
glfwDestroyWindow( window );
glfwTerminate();
return 0;
}
我系统上的代表输出:
cnt: 1884
0: 157288
1: 157288
2: 157289
3: 157288
4: 157287
5: 157286
6: 157292
7: 157286
8: 157289
9: 157288
相关文章:
- Mongodb c++驱动程序:如何查询元素的数组
- 删除一个线程上有数百万个字符串的大型哈希映射会影响另一个线程的性能
- 查询SQLite数据库中的日期
- OpenMP阵列性能较差
- 递归列出所有目录中的C++与Python与Ruby的性能
- 大小相等但成员数量不同的结构之间的性能差异
- 为什么constexpr的性能比正常表达式差
- 如何在ArangoDb AQL查询中指定数据库
- 在类中使用随机生成器时出现性能问题
- 在main()之外初始化std::vector会导致性能下降(多线程)
- Qt SQLite没有查询或参数计数不匹配
- 如何使用c++在VS 2019上运行SQL查询
- 从返回的顶点缓冲区查询顶点结构
- 查询性能计数器限制/加快幻灯片速度
- 检查从查询返回的任何行是否包含在字符串中的最高性能方法?
- 后续:异步屏幕外查询性能
- 查询性能计数器函数在在线判断中的使用
- 是否可以使用查询性能计数器(Win32,C)获取时间戳
- SQLite3 c++对5个双字段的SELECT查询的性能问题
- 查询性能指标休眠问题