Lucid Logix HYDRA tech brings together any GPUs for powerful matrimony

[Mr.BSEL]

http://www.engadget.com/2008/07/14/l...-powerful-mat/

If this is true, this could revolutionize the world of gpus.

[vitaminc]

was wondering why no one has posted this on XS news...

And btw, wasn't Lucid Logix one of the companies that NVidia has invested in? I remember the company was founded by a whole bunch of ex-Intel processor guys up in Oregon.

[BossBorot]

How will this chip be used assuming it works? Will the chip be an add in card or a add on chip to be put on mobos?

[Aberration]

Repost.

http://www.xtremesystems.org/forums/...d.php?t=128645

First posted 1/2007, and just recently revived.

[Xoulz]

Gulp!

I'm wondering if this isn't what INTEL was refering too, in some of their slides of the new X58
Another news source:[H] info...

Here is the Companies own website:
Lucid's HYDRA <--- 50 Patents


Pretty interesting stuff, specially when you consider the people who are employed by this company. Very impressive!




.

[BulldogPO]

Looks very promising indeed. Now just waiting for realworld applications.

[Donnie27]

Saw it last night. The only problem, is it something nVidia block via drivers, again! Remember, i975 didn't need anything to run SLI.

[bowman]

Hot on the heels of its new announcement of "Hydra", we caught up with Lucid's busy founder and President - and fired off a few questions to him regarding Hydra:

http://www.bootdaily.com/index.php?o...1153&Itemid=51

According to this podcast they're at IDF, with a stand. So maybe we'll know more.

[Kayin]

It's the ****ing Voltron of computers. Awesome.

[ferrari_freak]

Rofl I love how the first comment is "Ya, ya, but can it run Crysis?" Anyway, sounds like a really interesting technology, can't wait to see how this turns out.

[thephenom]

As much as that can theoretically work, it'll still require a certain amount of cooperation from NV and AMD for this to work.

[ColonelCain]

As much as that can theoretically work, it'll still require a certain amount of cooperation from NV and AMD for this to work.

That is what I would think, as well.

The only way in my mind that they could get this to realistically work, is to thoroughly benchmark EVERY semi-new GPU from both ATI and nVidia. Then, set certain paramaters in their "Hydra" engines' programing, to leverage the strengths of each GPU, and then use them accordingly.

Still, this looks damn near impossible.

[initialised]

How will this chip be used assuming it works? Will the chip be an add in card or a add on chip to be put on mobos?

The Lucid Chip sits on the motherboard to manage multiple GPUs as if they were a single resource. Alternatively it sits on a graphics card or daughter board on the PCIe bus and does the same thing.

So instead of the CPU/MCP seeing the lowest specced card (or only one with SLi) it sends data to the Lucid chip which passes it on to the GPUs according to their capabilities. Basically if you add a 4850 to 3870 at present the 4850 will be constrained by the R670. With Lucid managing data flow (perhaps adding latency) you would get performance scaling according to the capability of the card.

Basically it makes heterogeneous multiple GPUs work together in a way that ATi has been attempting and nVidia has been avoiding.

[Stumbo]

How will this chip be used assuming it works? Will the chip be an add in card or a add on chip to be put on mobos?

They website says it will be either, you will be able to purchase it as an add-on and some MBs will have it built in.

[Aberration]

Look at the picture and description. To me it appears that it wont be an add-on board, but added to the board of an x2 or something.

They state it has to be between the chipset and the GPU, which makes sense.

[Xoulz]

The Lucid Chip sits on the motherboard to manage multiple GPUs as if they were a single resource. Alternatively it sits on a graphics card or daughter board on the PCIe bus and does the same thing.

So instead of the CPU/MCP seeing the lowest specced card (or only one with SLi) it sends data to the Lucid chip which passes it on to the GPUs according to their capabilities. Basically if you add a 4850 to 3870 at present the 4850 will be constrained by the R670. With Lucid managing data flow (perhaps adding latency) you would get performance scaling according to the capability of the card.

Basically it makes heterogeneous multiple GPUs work together in a way that ATi has been attempting and nVidia has been avoiding.

+10

There is allot of technology (and patents) that went into making the HYDRA. SLI and Crossfire are dead-end technologies with the announcement of this chip.




.

[Aberration]

Well I wouldnt say SLI and CF are dead, they are just going to work better.

And 2009 is not that far away! Intel is invested in them. I wonder if they will buy em up and incorporate this into their chipsets. Boy that would be a strike for other chipset manufacturers.

[STEvil]

PSSST: 1999

http://www.firingsquad.com/hardware/...er/default.asp


Yeah....

[Hornet331]

PSSST: 1999

http://www.firingsquad.com/hardware/...er/default.asp


Yeah....

god, fs should have patented the names quad and tripple sli.

[Xoulz]

Well I wouldnt say SLI and CF are dead, they are just going to work better.

And 2009 is not that far away! Intel is invested in them. I wonder if they will buy em up and incorporate this into their chipsets. Boy that would be a strike for other chipset manufacturers.

Why do you say that?

Did you listen (or read) to one of the inventors explain the chip? There is no need for SLI or Crossfire, the HYDRA chip can do this natively with as many GPUs you can muster.

It's infinitely scalable and they hold all the patents on it. For SLI or Crossfire to get better, they would have to go down the same path, but cannot because of Lucid's 50 patents.

Just one look threw the Board members and I was sold, there are some heavyweights behind this chip.



.

[v_rr]

Why do you say that?

Did you listen (or read) to one of the inventors explain the chip? There is no need for SLI or Crossfire, the HYDRA chip can do this natively with as many GPUs you can muster.

It's infinitely scalable and they hold all the patents on it. For SLI or Crossfire to get better, they would have to go down the same path, but cannot because of Lucid's 50 patents.

Just one look threw the Board members and I was sold, there are some heavyweights behind this chip.



.

Have you seen this working or benchmarks of it?
No, so why the hype?

[Swatrecon_]

If this truly has promise, Nvidia will end up buying it so it doesn't bury SLI.

[Aberration]

Why do you say that?

Did you listen (or read) to one of the inventors explain the chip? There is no need for SLI or Crossfire, the HYDRA chip can do this natively with as many GPUs you can muster.

It's infinitely scalable and they hold all the patents on it. For SLI or Crossfire to get better, they would have to go down the same path, but cannot because of Lucid's 50 patents.

Just one look threw the Board members and I was sold, there are some heavyweights behind this chip.



.

*sigh*. Because it doesnt matter where this chip ends up, if its a multi-gpu solution from nV it will be marketted as SLI, from ATi it will be CF.

The marketting will just change, now it will just be "SLI with Hydra power!".

[Aberration]

If this truly has promise, Nvidia will end up buying it so it doesn't bury SLI.

Considering how Intel is invested, I doubt it.

[Xoulz]

*sigh*. Because it doesnt matter where this chip ends up, if its a multi-gpu solution from nV it will be marketted as SLI, from ATi it will be CF.

The marketting will just change, now it will just be "SLI with Hydra power!".

I see your point, but I think it is going to be more mianstream than that. Specialy if Mobo manufacturers start offering HYDRA boards. Then you can buy what ever graphic card (or cards) you want and just start plugging them in.

With no fear or worry if it is a Crossfire or SLI.


I'm still waiting to find out if those two technologies will even matter, since that engineer somewhat laughs when asked about it. I guess time will tell.




.