At IDF Lucid technology (heavily invested by Intel) demonstrated it's HYDRA technology. The idea is to "build a completely GPU-independent graphics scaling technology" which enables two to four completely different (Radeon / GeForce / models ) to work together in harmony with "little to no software overhead."
More info on this "Bridge' chip was presented at IDF. This might even be a revolution in the high-end gaming segment. Lucid is building a new "real-time distributed processing engine" system on a chip called HYDRA, which can mix and match any GPU from any manufacturer and work with any chipset, and piles it all together for performance scaling that Lucid claims is "near-linear" or even "above-linear."
Next problem that i see ... API calls, but exactly what does Hydra communicate to and with the GPUs? It can't send the hardware itself API calls, as these do not allow that. It must send machine code generated by the graphics driver. So all the difficult analysis of API calls and grouping of tasks and load balancing has to happen in software in the driver. And there where I see an issue. A lot of CPU overhead..
It's the above-linear part that particularly makes no sense though. My second thoughts on this topic is .. i have no idea how this will work out legally with companies as ATI and NVIDIA having patented their multi-GPU technology.
But even more importantly .. would you even like to bridge say two or three GeForce GTX 280 cards over an external solution rather than the original SLI or Crossfire method ?
Not being brainwashed at IDF, Guru3D would like to see some actual working hardware and a decent software implementations before making any real comments. And then do a side by side system performance comparison. As you can see, we here at Guru3D are a little sceptic about the product. As in the end, it'll also add more power consumption to the system and hey, again another level of additional costs, as you need to pay for yet another board in your PC.