Multi-brand GPU mayhem likely
ISRAELI OUTFIT LUCIDLOGIX Technologies is touting its brand-spanking-new Hydra Engine technology that will allow cross-branded GPUs to operate hand-in-hand. Sounds grand, doesn’t it?
This market-shaking feat is performed with the use of a dedicated SoC, dubbed the Hydra Engine, which distributes processing tasks to different GPUs and optimises data flow to and from each one, maximising (in theory) the potential performance per GPU. The firm uses some nifty data compression/decompression algorithms to simplify things, it says, in a cost-effective manner (although, as we all know, GPU economics reside in the Twilight Zone).
The claim is quite grand, as it states that this will allow you to throw a bunch of different graphics cards into the mix and get “near-linear to above-linear performance” across different GPUs. The Hydra Engine analyses and adapts on-the-fly to feed each GPU, creating parallel processing engines under any CPU, chipset and GPU combo.
A simple graph on the site shows several graphics cards wearing green and red colours (nudge nudge wink wink) fed by the Hydra Engine and outputting to a display. It’s pretty vague, if you ask us, especially since they don’t mention if you can combine the processing power of the GPUs.
As the technology is described as a “real time distributed processing engine”, and if their idea of distribution is anywhere close to CPU load balancing, it’s likely to be just distributing processes to the idlest GPU available. We don’t expect Hydra to SLI/Crossfire ATI and Nvidia GPUs into a single graphics processing behemoth, but rather keep things nice and tidy within the data path.
We called them up to get some clarification, but the prez was already on the phone with another inquisitive hack and was getting back to us “as soon as possible”. We’re waiting for that call.
In the meantime you can see that parallel processing is becoming big, both on CPUs and GPUs, and that a device of this nature could do a lot more than just improve “gaming” and “office computing”. We’re sure render farms would be more than happy to tap into this technology... and we’re quite sure there’s even a place for the IP within Intel’s own Larrabee or AMD’s Fusion.
Getting back to the details: the Hydra Engine can be soldered onto the motherboard or slotted as an add-in board, depending on what each partner will want to do. The technology will support OpenGL and DirectX and should be available through a number of said partners in early 2009.
If the firm isn’t gobbled up in the meantime, that is. Guess which chip maker has a stake through the company right now. Yeah, the one beginning with I. µ
L’Inq
LucidLogix
Share this:
ShareBusiness Intelligence Software
Customer Relationship Management (CRM)
Enterprise Accounting Software
Enterprise Resource Planning (ERP)
Enterprise Systems Management
"...above-linear performance..."

Nice trick, if true, which I strongly doubt. That'd be like claiming you found a way so 1 + 1 = 2.2. Sorry, can't buy it. If true, it means that both ATI and Nvidia don't know how to utilize their own chips properly to give maximal performance.
Being an Israeli outfit a big chip company beginning with the letter ‘I’ having a steak in Lucidlogix would be the most likely candidate. Especially since most of the latest and greatest of ‘I’s tech comes from Israel.
If it works, I'll take one. I'll add a GTX 280 to my HD 4870.

So if I understood correctly, basically it won't use the second card until the need arises?
Only some idiotic VC would fund something this dumb. A total crock of you know what. If there were 20 graphics asic builders, this could be interesting, but there's...wait let me think about it.... TWO

Larrabee shmarrabee... total waste of effort. Solve the energy crisis instead
In theory when you have 3 cards they all could be doing some overhead thinking they are alone which only one really has to take care of and then you can have 1+1+1=3.2.
It IS possible, in theory.
An example, let's say you have a busy restaurant with 3 cooks, now they need to get some stuff from the freezer, one cook can go and bring supplies for all 3 cooks rather than each going, that means while he's doing that the other 2 keep working and they are 'scaling above linear'.
"...above-linear performance..." 

Rich Wargo, above-linear performance is possible... in theory. The flaw with your mathematically analogy is that you are assume all bits are integers or the same type of data. Let's say NVIDIA's cards are better at doing A and ATI's cards are better at doing B. Given a workload mixed of A and B, a NVIDIA and ATI card would perform better than just two of either. Hence, above linear performance... assuming there is not a lot overhead for the scheduler/manager.
IT really is possible, if you had the software available you could send the best possible data to the best possible hardware. ATI hardware right now when giving data it can work on in paralel is ludicrously faster than Nvidia cards, likewise, when given data that can only be done in order then Nvidia cards are much faster. WIth a mix on a single card neither card is performance at 100% maximum capacity. If you could split up all available calculations to the right gpu you definately could get more than linear performance. However the likelyhood of being able to optimise software to do this efficiently at the speed its needed on the fly is incredibly small. If someone pulled it off they'd be simply put, one of the best coders/minds around, I very much doubt anyone could make it work well.
parallel processing is becoming big, both on CPUs and GPU.
Word BIG Means Huge. With Microsofts promised Multi Core O/S Next, Executionares Eye Body for musculature strengths & Migon is code or actually FX MultiCore.. 

Todays Core 4,3,2,1, World just is Old Meat, SoleFood. Or So Next Show of Microsofts' Research Should show & Convience Multi Seekers Crowding to Go Hair Brainered. 64 Cores is that enough of Step & Soon Enough GPUs* CpGPU* & other discrete PU*s.
So Plan On Some Strong demo(Comxces) hardware in 18 Months, too. By Then 2 Billion transistor will be in Mid glory & 64 Cores, Smokin' Joe States, Welll, hody 'bout Eight billie tranies?, fits curve don't Et? 
writing never changes, just numbers.
That Shurely Bee 64 Bit holy with ?Nero Support or ActiveX written or how about little IE 64 thats works tad bit more often?By Then, Think About It!!

Maybe its coming to table, yet when will second tier go $ that parallel? Actual software NOT That Much More to control multi core FX enviorment ( about dozen pins NEW Inst?) & Crossfire Done Multi mfgs Cards in Crossfire X in 590 world with ease. Obviously, just faster processor is NOT Solution, Better Software updating IS.Even Today.

Ultie States. "Bring On BIG Stuff, I'm Ready, Willing & Able.."..So Where Is It?
SPIES:LOAD cameras.....
* meaning Reserved
TS Drashek
Are you sure this companies name isn't Lucifer-Logix ? ! ? !
Pigs will fly, Glaze3D will be in stores by Christmas and M$ will go open source.

No need to read all above, comment title says it all.....

Super-linear speed-ups by combining different architectures is quite plausible.

For example, lets say Nvidia GPUs rock at 1D FFTs but suck at 2D FFTs while ATI GPUs suck at 1D FFTs and rock at 2D FFTs.

If you have a workload that combines both 1D and 2D FFTs, then neither Nvidia nor ATI will be optimal. But if you combine them both and are smart enough to farm out all the 1D FFTs to the Nvidia GPU and all the 2D FFTs to the ATI GPU you could easily see aggregate performance increase by more than 2x,

It all depends on the workload and the architecture of the systems involved.
Just this morning (early tee tyme) I was thinking about this. It all takes a bit to sink in, but I thought I remembered reading that Larrabee would work with crossfire. Well after thinking that the new Chipxiillas required greater than 16 level PCBs (a muchroom piazza for all the strolls down memory arcade), I concluded that Intel must be in it for the money. Whose to say bobs did his work, eh? Who's indeed? AIB quite the marrying kind. Make mine wit a jimmy of hot relish, wouldya ducks? I won't 404not fifth-wheeling that Anjerlener motherbrood, now that the news about the twins.
No. really!
So when both different graphics card have slightly different precision and rasterisation pipelines you will get nice seams where the tiles join up?

Smells like BS to me..
To do even half of what is necessary to enable this sort of unity it would require the licensing of so many different technologies from both AMD/Nvidia and probably Intel as well, and we also know theres no way in high fking hell that the above 3 companies are going to license said technologies for this purpose

The best you can hope for is some way of doing a dynamic recompile of code for said graphics units, be it for CUDA or AMD's offering as the case may be. And the processing power required to do that for normal graphics tasks would be suicide and your be hoping for a less than linear improvement at best