• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Lucid can scale with 2 GPUs close to linear

Permabanned
Joined
11 Jan 2009
Posts
252
http://www.fudzilla.com/index.php?option=com_content&task=view&id=11408&Itemid=1
We had an interesting meeting with the president and CIO of Lucid, a company that gained a lot of attention at IDF 2008. Lucid promises a small chip that can make your two GPU scale almost linear. After a long discussion we learned that Linear is almost linear, and that linear is best case scenario and that real world performance should be close to linear.

This would mean that two GPUs, let's say in SLI or Crossfire, powered by Lucid chip and driver, could score close to 100 percent faster than a single GPU. Let's say that if a single GPU would score 100 FPS in one game, two would score close to 200FPS.

The chips that will find their place in both motherboards and graphics cards will appear on the market in 2H 2009. The prototypes are already finished, but the company didn't want to talk about any specific partners.

Anyone who has a PLX chip that makes two cards run together could be Lucid’s potential customer, and obviously some guys are interested. This leaves a lot of unclear questions to be answered, as we suspect that both ATI and Nvidia might have something against a chip that makes SLI and Crossfire look like child's play, and that this could lead to a driver that will block Lucid’s ability to shine.

We don’t want to talk about moral dilemmas, but we suspect that this might be the case. Lucid promised to show us a demo in the near future and we believe that Intel is keeping an eye on Lucid for its Larrabee school science project.
 
Very interesting multi GPU systems may start to live upto the marketing hype now ! .... but I bet they get bought out and never heard of again.
 
Let's say that if a single GPU would score 100 FPS in one game, two would score close to 200FPS.

but that's just not plausable - the 100fps score isn't just down to the 1st GPU, it's also affected by CPU, ram and whatnot.. even linear gpu multicore performance won't give you double the fps, without also doubling the rest of the system's power. It's like saying the 2nd graphics card is as powerful as the 1st graphics card plus CPU.
 
uvmain, some CF titles scale 100%

I can understand completely grfx bottlenecked games scaling that well.. say, Crysis etc.

*warning, arbitrary, ideological numbers ahead*

But let's say for example you're running a game at it's perfectly matched resolution, so that's it's neither CPU nor GPU bottlenecked. The CPU and GPU would be lending themselves to 50% of the performance each. Add another GPU, and you're adding another '50%'. 50% times three equals 150%, not 200%. In decently coded games that aren't completely GPU bottlenecked, you just can't expect double the FPS by chucking double the GPU's at them.
 
well in that case, what exactly is the point of this 'Lucid'? It reports to give 'almost linear performance', but apparantly we've already got that?

Ah, but apparently it works with more than two GPUs, again, in a linear way. Tri/Quad SLi and 3/4-way Crossfire X provide nothing in that sort of capacity.
 
awpc, hope ur joking
Nope. Whats the big deal here. X58 platform benchmarks have shown that both SLI & Crossfire achieve very close to double the FPS of a single card. By my maths that's almost 100% give or take a % or 2. And AFAIK that means each card is almost giving 100% performance taking into account the loses for OS & driver overheads. Does anyone really think that ATI/Nvidia are just going to let some 3rd party take away part of their core business by improving older model performance and reducing the demand for newer or multi gpu cards just like that. All they have to do is block it @ the driver level like they did with SLI for years. ATI also have the sideport to enable on X2 cards when/if they feel the need to do so.

Maybe I am missing something here and someone could explain what that is!!
 
Well, if they've invented something pointless here as many of you suggest, maybe they should work on a chip that stops 3/4 of my Q6600 picking it's flaming nose most of the time ;)
 
Nope. Whats the big deal here. X58 platform benchmarks have shown that both SLI & Crossfire achieve very close to double the FPS of a single card. By my maths that's almost 100% give or take a % or 2. And AFAIK that means each card is almost giving 100% performance taking into account the loses for OS & driver overheads. Does anyone really think that ATI/Nvidia are just going to let some 3rd party take away part of their core business by improving older model performance and reducing the demand for newer or multi gpu cards just like that. All they have to do is block it @ the driver level like they did with SLI for years. ATI also have the sideport to enable on X2 cards when/if they feel the need to do so.

Maybe I am missing something here and someone could explain what that is!!



wana link me to benches that prove that?


remember with this chip u can mix and match nivdia and ati cards
 
remember with this chip u can mix and match nivdia and ati cards

From The Inq's report it states "any 4 cards from a given manufacturer".

I assume this is down to Windows and (afaik) the issue of multiple graphics drivers.

I would love to be proven wrong and be able to have my old 8800GTX alongside my 4870x2 :D
 
techreport article and Anandtech article

Hopefully this will clear up some of the confusion on how it can achieve linear scaling

techreport said:
The Hydra 100 then appears to the host OS as a PCIe device, with its own driver. It intercepts calls made to the most common graphics APIs—OpenGL, DirectX 9/10/10.1—and reads in all of the calls required to draw an entire frame of imagery. Lucid's driver and the Hydra 100's RISC logic then collaborate on breaking down all of the work required to produce that frame, dividing the work required into tasks, determining where the bottlenecks will likely be for this particular frame, and assigning the tasks to the available rendering resources (two or more GPUs) in real time—for graphics, that's within the span of milliseconds. The GPUs then complete the work assigned to them and return the results to the Hydra 100 via PCI Express. The Hydra streams in the images from the GPUs, combines them as appropriate via its compositing engine, and streams the results back to the GPU connected to the monitor for display.
 
Back
Top Bottom