cuda array to opengl texture

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi!
Is there a easy way to copy cuda array to opengl texture by one function without:
cudaGraphicsMapResources → cudaGraphicsSubResourceGetMappedArray → copy → UnmapResources ??

P.S. The reason is next: in each frame-iteration of raycasting I map for picture, tnen I create two threads for two GPUs for raycasting (half on screen on each of them) and then I unmap Ouputput. In this case somehow map/unmap takes a loooot of time and extremely downgrade fps. I didn`t find any explanation of this.

2

You can map CUDA memory to pixel buffer object.

3

Thats not really what I asked. I don`t want to use cudaMap/unmap at all. Just to copy somehow directly to opengl texture from cudaArray if possible without getting pointer by cuda.
Or of course if somebody knows something about drop of fps in this situation…

4

Then I dunno is it possible to write to GL without mapping…

Is it essential to use two GPUs for this task? Maybe this is the problem? I mean wrong utilisation of the GPUs…

MK