Lucid Hydra 200 - It's almost here!!!

why the heck does everyone think nvidia is the devil and eats babies, while ati is your best friend and would bake you a cake if you wanted one? if (huge if) this works ati is just as likely to block this as nvidia is.

because AMD is on the ropes, and that makes people willing to believe AMD might genuinely try to ingratiate themselves with consumers rather than putting out nothing but marketing fluff.

i can't speak for everyone, but i think all businesses are the devil and eat babies, and AMD would try to screw me out of every dollar i have if they only had the means.

however, i think nvidia in particular maintains a rather obnoxious public image akin to a adolescent male, mainly because that's the group of people (mentally at least) they want to keep close. even when they release products that lag behind in some way (whether that's raw performance, bang/buck, efficiency or whatever), they know that acting like a bro earns them some loyalty from the "OMFG i pwned ur mom" crowd. and the thing is, it works.
 
And if and when it does appear, it will take Nvidia about 3 seconds to push out a block in their drivers.

They have said officially now that they will embrace it if it's better. I'm guessing they make more off of video cards then the nforce mobo's.
 
ATI & Nvidia couldn't buy Lucid because Intel did

Incorrect. Intel Capital (related to but separate from Intel) invested in them. There's a huge difference between a company trying to invest to see returns (like a portfolio) and a company purchasing another.
 
You guys do realize that this would relieve a lot of burden for the programmers behind the drivers. It would reduce costs for both nvidia and amd. Open up new opportunities for both also. A person can have a 5870 and also have the latest nvidia card working together. New opportunities to convert each others sides to buy their products. That would suck though if physx is still blocked though. I think they'll remove that though if this turns into a hit.
 
You guys do realize that this would relieve a lot of burden for the programmers behind the drivers. It would reduce costs for both nvidia and amd. Open up new opportunities for both also. A person can have a 5870 and also have the latest nvidia card working together. New opportunities to convert each others sides to buy their products. That would suck though if physx is still blocked though. I think they'll remove that though if this turns into a hit.

wow, looks like you are drinking the kool aid AND someone put ecstacy in it.

you are crazy to think this would reduce costs. its hard enough for both companies to make working video drivers, do you think its going to be easier when they would have to consider what happens when a video card from a a different vendor with different capabilities is also in the pc and rendering different parts of a scene every single frame?
Posted via [H] Mobile Device
 
Lets see .. a product that completely cancels out nearly half the marketing of high end video cards (mid-range as well) and allows cross competitor mix matching with nearly a 90 precent (claimed) speed boast in many circumstances ...

And they don't think Nvidia/ATI will simply code out the support for this? I mean it completely slams SLI/Crossfire in the face and people expect Nvidia/ATI to not fight against it?

Even if it becomes a reality and is something I can buy I see both Nvidia and ATI doing all they can to block this hardware from working.
 
Do you really think NVIDIA and AMD like having to profile all new apps, and avoid introducing corruption?

Profiling is not the perfect solution, but it is the best out there today. People can talk up this algorithmic solution but at the end of the day there are a ton of shortcomings, and I have major doubts about it scaling well, not to mention the potential driver/API messes it can cause.

But the bottom line is, don't you think that a couple of multi-billion dollar companies would've already looked into this approach? If it had panned out as something that might work and alleviate the pains of profiling, don't you think they would've done it by now?

I'm pretty sure there's a reason we've only seen this demo'd on like 3 different games over the years.
 
You guys do realize that this would relieve a lot of burden for the programmers behind the drivers. It would reduce costs for both nvidia and amd. Open up new opportunities for both also. A person can have a 5870 and also have the latest nvidia card working together. New opportunities to convert each others sides to buy their products. That would suck though if physx is still blocked though. I think they'll remove that though if this turns into a hit.

Every generation one side has to lose. That is great and all, so long as you aren't on the team that stands to lose market share... so it will always have an enemy.
 
But the bottom line is, don't you think that a couple of multi-billion dollar companies would've already looked into this approach? If it had panned out as something that might work... don't you think they would've done it by now?

there were several years where the same thing could have been said about intel, the undisputed giant, regarding integrated memory controllers.

there's a pretty big difference between "looking into" an approach in the abstract (on paper) and actually creating a working engineering sample. when brainstorming multi-GPU technology, i have to imagine NV and ATI would have considered several options and picked the one with the least financial risk, not the one that seemed like the most elegant and effective solution.i don't think it's too hard to imagine a converstaion between NV or ATI management and engineers going something like...

engineer: "we have the capability of making 2 of our GPUs work in unison, so we can sell twice as many, and we've got these two possible methods. we can make hardware that'll do it, but it will probably take a while to develop the necessary technology and iron out all the bugs. in the end it might work better, but we don't know for sure how much time and money it would take to develop."

executive: "what's the other method?"

engineer: "well, we can program the drivers to do it. we won't need to develop new hardware, and we can always iron out bugs in subsequent driver revisions. we can have something that sort of works to the market pretty quickly, but-"

executive: "that's what we're doing."

just because a particular method might have been risky doesn't mean it's inferior, and engineers don't get the final say in what technology gets developed and what doesn't. i'm not saying i blindly believe this chip will be better than SLI or CF, only that i'll wait to see comparisons of it against SLI and CF before jumping to conclusions.
 
Lets see .. a product that completely cancels out nearly half the marketing of high end video cards (mid-range as well) and allows cross competitor mix matching with nearly a 90 precent (claimed) speed boast in many circumstances ...

And they don't think Nvidia/ATI will simply code out the support for this? I mean it completely slams SLI/Crossfire in the face and people expect Nvidia/ATI to not fight against it?

Even if it becomes a reality and is something I can buy I see both Nvidia and ATI doing all they can to block this hardware from working.

THIS..

/ THREAD
 
jesus rtft people

Nvidia has officially stated now that if it works they will embrace it, so they can sell more video cards.
 
10/08/09 http://www.extremetech.com/article2/0,2845,2353999,00.asp Not much new info here, good filler none the less.

"Working in Lucid's favor is Intel's investment in the company. This has led some to wonder if Intel's decision to remove the Northbridge controller chip from the Bloomfield processor line when creating the Lynnfield line had something to do with the Hydra. Shifting the PCIe processing elements off of the motherboard and onto the chip opened up significant space for PCIe slots, an architectural change that suits the Hydra 200's needs and product release timing."

Interesting...
 
Lets see .. a product that completely cancels out nearly half the marketing of high end video cards (mid-range as well) and allows cross competitor mix matching with nearly a 90 precent (claimed) speed boast in many circumstances ...

And they don't think Nvidia/ATI will simply code out the support for this? I mean it completely slams SLI/Crossfire in the face and people expect Nvidia/ATI to not fight against it?

Even if it becomes a reality and is something I can buy I see both Nvidia and ATI doing all they can to block this hardware from working.

Huh?

Say I have a 4870, and AMD has their new 5870 out, and Nvidia has the GT300 out, and I want some more GPU power so decide to buy a new GPU.

If AMD say for example support Hydra (or more specifically don't hinder its usage) then great I can grab a 5870 and use it along side my 4870 and enjoy the benefit of having the additional speed compared to just binning, selling my 4870.

If Nvidia don't support Hydra then getting a GT300 card is less attractive because I have to ditch my 4870 and I only get the performance of the GT300.

Nvidia and AMD are in the business of building and selling GPU's, by limiting Hydra they simply limit the circumstances in which their GPUs can be sold, that's bad for business.
 
Huh?

Say I have a 4870, and AMD has their new 5870 out, and Nvidia has the GT300 out, and I want some more GPU power so decide to buy a new GPU.

If AMD say for example support Hydra (or more specifically don't hinder its usage) then great I can grab a 5870 and use it along side my 4870 and enjoy the benefit of having the additional speed compared to just binning, selling my 4870.

If Nvidia don't support Hydra then getting a GT300 card is less attractive because I have to ditch my 4870 and I only get the performance of the GT300.

Nvidia and AMD are in the business of building and selling GPU's, by limiting Hydra they simply limit the circumstances in which their GPUs can be sold, that's bad for business.

in this example, amd would rather have you buy a 5890/5870x2 or a second 5870. gpus companies make less money if you can leverage old gpus for a performance increase. plus this solution would not work for dx11 games.
Posted via [H] Mobile Device
 
This product is completely pointless. Why would you buy an Nvidia and ATI gpu to do this? If you're going to buy a new gpu then buy another one of the same brand. And if you want to make the argument "i can't afford the exact same gpu", you can't tell me that they will take a low end and high end gpu and put them together for a significant performance boost over just the high end GPU alone. I wouldn't believe you.
 
I smell fail.

Please don't shit in my lawn, you I.A..

This product is completely pointless. Why would you buy an Nvidia and ATI gpu to do this? If you're going to buy a new gpu then buy another one of the same brand. And if you want to make the argument "i can't afford the exact same gpu", you can't tell me that they will take a low end and high end gpu and put them together for a significant performance boost over just the high end GPU alone. I wouldn't believe you.

Tell me, how much did you bother reading about this technology, bigdog, you fine fellow [H]ardforum member?
 
Yep the "owner" of Lucid Hydra get pissed because i don't trust in its future how cute! Can't wait to see the results and laugh ;)
 
This product is completely pointless.
You've totally missed the point.

This isn't intended as a workaround for the limitations of Crossfire and SLI. The fact that you can mix different models, different series and different brands is incidental. What they're trying to do is provide better workload partitioning with no requirement for software support.

If it turns out to be slower than CF or SLI, then it'll be pointless :D .
 
You've totally missed the point.

This isn't intended as a workaround for the limitations of Crossfire and SLI. The fact that you can mix different models, different series and different brands is incidental. What they're trying to do is provide better workload partitioning with no requirement for software support.

If it turns out to be slower than CF or SLI, then it'll be pointless :D .

QFT

I'm watching this intently, if it turns out to offer what it seeks to - I'll likely build a new rig.
 
When an i7 model becomes available you better believe im going to get that and get another video card because this makes multi-gpu seem more than reasonable
 
My wife is gonna hate this, now she'll get two generations old video cards when it's her turn to "upgrade" ;)
 
Can't wait to see benchmarks + game compatibility. The concept is brilliant and obviously they have working hardware but to me, this would only be good if it was faster than SLI/Crossfire. Not really interested in mixing old/new cards from either the same/different vendors. I would much rather buy newer DX11 cards and see whether the Hydra was faster than ATI's CF or Nvidia's SLI.
 
That's the entire point, Hydra is supposed to offer 90% (or better) scaling, where as SLI and crossfire hover between 40% and 60% depending upon the game.
 
Uh, where are you getting that FUD? It is not supposed to offer better scaling. Even the company does not claim that and will not release performance figures, saying only that it will be "competitive."

The "entire point" is the mixed GPU bit.
 
You've totally missed the point.

This isn't intended as a workaround for the limitations of Crossfire and SLI. The fact that you can mix different models, different series and different brands is incidental. What they're trying to do is provide better workload partitioning with no requirement for software support.

If it turns out to be slower than CF or SLI, then it'll be pointless :D .
... I'll give you that. Thanks.
 
Uh, where are you getting that FUD? It is not supposed to offer better scaling. Even the company does not claim that and will not release performance figures, saying only that it will be "competitive."

The "entire point" is the mixed GPU bit.

actually no
 
Back
Top