Hi 
there are two parts in what you are trying to do I think. One is generating meshes. Two is to rendering meshes. (And step three might be moving or deforming the mesh)
The thing that slows down frame rate is sending lots of data to the GPU on every single frame.
By keeping the generated mesh data in the GPU the rendering is much faster than when you send a new mesh to the GPU on every frame. If you are generating a heavy new mesh on every frame it will not be fast, even if you use low level OpenGL. But you can still deform big meshes in the GPU with shaders, which can be fast because you don’t need to upload all the vertices again and again, only a few values to deform what you have.
What are you trying to do? Do you have an image to share?