Like the idea of running a gaming rig with multiple GPU’s a-crankin’ but don’t want to be tied down to one chipset or graphics card manufacturer? Well Lucid’s gotcha covered. Enter the Hydra 200 real time distributed processing engine.
When first announced over a year ago many people, myself included, expected nothing. A chip that would distribute the graphics process to any GPU regardless of manufacturer and without proprietary cables and connectors, without performance sacrifices? Ok, sure. Well here we are about a year later and the Hydra 200 is about to make us eat our words.
The Hydra 200 chip is a small (18-23mm) chip that uses 6W of power and can be placed on a motherboard or expansion card to handle the distribution of the graphics processing requirements to the cards you have installed. Since it’s a separate chip, it doesn’t care what card that might be. AMD and Nvidia’s SLI concoctions require you to have a specific chipset and then use a special connector to get multiple cards to work in conjunction. Both graphics card manufacturers then use software in the form of drivers to distribute the graphics workload among the GPU’s. With Lucid’s chip the software is on the silicone and handles the processing before it gets to the cards so all the cards need to think about is pumping out the results of the work it gets.
I’m still not completely sold on it using “100%” of the performance of every card attached but it is intriguing. The Hydra is Vista and Windows 7 compatible and supports DirectX’s 9c and 10.1 as well as being DirectX 11 ready. MSI appears to be the first one out of the gate using a Hydra 200 chip on their “Big Bang” motherboards due out on October 29th. Hopefully we’ll get some performance benchmarks soon after that.
Comments