c++带有缓冲区资源的 OpenGL 多线程

发布于 2024-11-09 16:15:26 字数 627 浏览 0 评论 0原文

我有一个 OpenGL 程序需要定期更新纹理。但同时我希望程序在更新这些纹理时能够做出响应（具体来说，继续运行绘制/显示代码）。

但这似乎是不可能的：如果我让 thread1 执行绘制/显示代码，而 thread2 执行纹理移动，那么它们将在资源 GL_ARRAY_BUFFER_ARB 下发生冲突，因为 thread2 必须保留该资源才能将某些纹理移动到 vbo。我需要 GL_ARRAY_BUFFER_ARB 来为 thread1 执行绘制/显示代码，因为它使用不同的 vbo。

例如，此代码

glBindBufferARB(GL_ARRAY_BUFFER_ARB, tVboId);
    glBufferDataARB(GL_ARRAY_BUFFER_ARB, numVertices*2*sizeof(GLfloat), texCoords, GL_DYNAMIC_DRAW_ARB);  
    glBindBufferARB(GL_ARRAY_BUFFER_ARB,0);

将移动一些纹理，但这需要一段时间。在此期间，显示代码应该运行多次，但它会崩溃，因为 GL_ARRAY_BUFFER_ARB 正在使用中。

我以为我可以做类似 GL_ARRAY_BUFFER_ARB2 的事情，但我认为没有这样的事情。

原文

I have an OpenGL program that needs to periodically update the textures. But at the same time I want the program to be responsive (specifically, to continue running the draw/display code) while it's updating these textures.

But this seems impossible: If I make thread1 do the draw/display code and thread2 to the texture moving, then they will conflict under the resource GL_ARRAY_BUFFER_ARB because thread2 has to keep that resource to move some textures over to a vbo. And I need GL_ARRAY_BUFFER_ARB to do the draw/display code for thread1 because that uses different vbo's.

For example this code

glBindBufferARB(GL_ARRAY_BUFFER_ARB, tVboId);
    glBufferDataARB(GL_ARRAY_BUFFER_ARB, numVertices*2*sizeof(GLfloat), texCoords, GL_DYNAMIC_DRAW_ARB);  
    glBindBufferARB(GL_ARRAY_BUFFER_ARB,0);

will move some textures over but it will take a while. During that time the display code is supposed to run many times but it will instead crash, because GL_ARRAY_BUFFER_ARB is in use.

I thought I could do something like GL_ARRAY_BUFFER_ARB2 but there is no such thing, I think.

分享到QQ

分享到微博