HYDRA is a dedicated silicon with sole purpose of scaling GPUs. Though there is no graphics processing logic on the HYDRA chip, what the chip can do is redistribute graphics workloads across multiple GPUs in real-time. The HYDRA technology also includes a unique software driver that rests between the DirectX architecture and the GPU vendor driver.
(hopefully this comes to OpenGL, but still it is really cool)
Here is an Idea of the difference:
Normal GPUs in SLI or Crossfire take all the work and each one does the work evenly. That wastes the cards. Hydra splits the workload like so
GPU does lighting and contrast and some other things.
GPU 2 does darkening and texture, etc..
They both do different work loads then the two seperate images are put back together and you get a better image. The performance increase is like 15-40%