Vigile writes "Last year a small company called Lucid promised us GPU scaling across multiple GPU generations with near-linear performance gains without restrictions of SLI or CrossFire. The company has been silent for some time, but now it is not only ready to demonstrate the 2nd generation hardware, but also to show the first retail product that will be available with HYDRA technology. In this article there is a quick look at the MSI 'Big Bang' motherboard that sports the P55 chipset and HYDRA chip and also shows some demos of AMD HD 4890 and NVIDIA GTX 260 graphics cards working together for game rendering. Truly platform-independent GPU scaling is nearly here and the flexibility it will offer gamers could be impressive."
If it is essentially just a load-balancer, why can't it be done in software?
The article only mentions DirectX, no word about OpenGL, so it must be not a pure hardware solution. If all it does is re-routing of D3D calls, why CPU can't do it?
That would require CPU. Rendering a game at 60Hz not only requires a GPU that can render your imagery within 16.7ms, but also requires the software running on the CPU to issue its directx/openGL commands in 16.7ms.
The idea here is to improve performance. There are many complex calculations that need to be performed each frame to properly load balance the cards to achieve significant performance gains. If all that work was being performed by the CPU, it's quite possible that the rendering process could become CPU limited while the graphics cards sit waiting for the CPU to decide which card does what.
Yes VirtualGL can do this easily.
it can attach to any window also and then use GPUS on multiple machines at once.
Its very stable and easy to use also
Forget about performance, what about power saving? For a couple of years now we have been promised the ability to shut down a graphics card and rely just on the on-board chip for desktop use, with the card kicking in when a game is launched. No-one seems to have actually implemented it yet though.
Forget about performance, what about power saving? For a couple of years now we have been promised the ability to shut down a graphics card and rely just on the on-board chip for desktop use, with the card kicking in when a game is launched. No-one seems to have actually implemented it yet though.
I know nVidia does, and I believe ATI has similar technology. There's a weak GPU onboard the chipset, but it can then switch to the faster offboard GPU when you want the grunt (at the expense of battery life).
Of course it's only supported on laptops. Laptops are where there is a solid business case for the feature.
I'm sure it will show up on desktop systems when it becomes more expensive to not do it.
Getting full screen flash to perform well on Linux is not the fault of your hardware, which is what this solution is. I have quite a mediocre PC running Windows, and Flash works full screen just fine.
I also have a gaming PC with a more than capable second card not being used, but which would probably allow me the small performance boost I need to keep me from upgrading just yet. I think that I'm more the target market than you are.
Fullscreen flash on Linux is just one of those/. memes that gets recycled past its "best by" date. It probably still applies to some combinations of graphics hardware and driver, especially older stuff.
Just FYI, I picked up a pair of Radeon 4670's for ~$50 each open-box from Newegg and a bridge for $7, and in Crossfire they perform about like a 4850 from my benchmarking and basic testing. And they don't need a secondary power cable, either. It's a cheap, easy upgrade if you have two x8/x16 PCIe slots.
Finally, we can have asynchronous GPU pairing? And cross-brands to boot? What's incredible is having heard nothing about this for so long, TFA now says a product may hit the market in the next 30 days.
I take it that by sidestepping Crossfire and SLI technology, they won't have to pay any licensing fees to either AMD or NVIDIA. Hopefully the patent trolls won't be able to find any fodder that would prevent and delay commercial release.
I think NVIDIA has some sort of asymmetrical SLI mode available on its mobos with built-in video cards. It allows the weak built in card to help a little bit with the big video card installed in the main PCI-E slot.
10% isn't a big deal? There are people who go to crazy extremes just to tweak out an extra 1-3% with entire sub markets dedicated to them, so yeah 10% is worth it.
To accompany this ability to intelligently divide up the graphics workload, Lucid is offering up scaling between GPUs of any KIND within a brand (only ATI with ATI, NVIDIA with NVIDIA) and the ability to load balance GPUs based on performance and other criteria.
So what is the deal? Is it cross-brand or not?
Also, they are only planning to launch their chipset on one motherboard with one manufacturer. It all sounds like a short-lived gimmick to me.
From the article, "The HYDRA technology, as it is called, is a combination of hardware (in the form of a dedicated chip) and software (a driver that sits between the OS and DirectX)", I can't wait for this software technology to be available for GNU/Linux. But.. something tells me it will take a while as never ATI and NVidia chips can not even do 3D using free software as of today and support seems to be years away. And yes, I know, there is some unstable proprietary binary blob available for my ATI card wh
So you can't wait until approximately...20never? You won't see drivers for anything like this in Linux. You'll be lucky to get decent bog standard 3D drivers.
"Immoral"? What, because it's proprietary? Are you serious? Get ready to throw out your whole computer, because the whole damn thing is proprietary. You don't have circuit diagrams for the cpu or gpu, you don't have firmware code, nothing. Before you start taking the "moral" high ground about proprietary components, look at what you're typing on. There's plenty of room in the world for proprietary and open source to coexist, RMS' rantings not withstanding.
Keep an eye on the radeon development. They just pushed OpenGL 1.4 acceleration to the Radeon driver for all ATI cards (including the current r600/700 cards, the 2xxx/3xxx/4xxx series), and it's just getting better. It'll really fly once Gallium3D drops and allows GLSL and other improvements. Most distros should be including it when they get the 2.6.32 kernels shipping with them. So, Fedora 12 alphas have it running, Ubuntu 9.10 should have 3D without the KMS for the new radeons, and things in general are j
I hope you die young. Seriously. If we get world hunger solved, and peace eternal, people will start to complain about even less important stuff. People complain about things, it's part of human nature. Just because 500 people died in Africa today before I got out of bed doesn't mean I don't feel that particular idiot at work is a friggin' [censored].
You can't deny people's feelings with a rational appeal to global standards.
We live in a world where thousands of children starve to death every day, people are killed or imprisoned for expressing their beliefs, women/minorities/everybody are oppressed, and few people really care about any of it, because it's all someone else's problem. I find it kind of funny (and more than a little sad) that the use of a driver can be blithely written off as "immoral" just because you can't download the source.
Some people rape children. How can you possibly think shoplifting is immoral?
Cute, you are trying to pull an common didactic trick by sidestepping the issue with overexerting another. I know grade schoolers like to do it:
Kid: I don't want to do my homework, it's so stupid. Parent: For the last time kid, do your homework! Kid: You know that hurricane Jane killed 5000 people yesterday on the west cost? And you are upset about some measly homework? How can you, there are so many worse problems on the world. Parent: [...]
There are going to be some performance hits compared to native crossfile/sli implementations. There are three models of the Hydra 200 part, and they each differ in their pcie lanes. The high-end model, which is going on the MSI motherboard, sports two x16 pcie lanes from the chip to the graphics cards (configurable as 2x16, 1x16 + 2x8, or 4x8), but only a single x16 lane from the chip to the pcie controller. So, where a good high-end crossfire or sli board will have two x16 pcie lanes from the controller to the slots for the gpus, this solution will be limited to one x16, limiting the bandwidth available to each graphics card. Exactly how much of a performance hit this would incur remains to be seen, and it probably depends on the cards being used (an older 8000 series geforce doesn't need/won't use as much bandwidth as a gtx 295, for example), but I would expect as gpus grow more powerful and require more bandwidth to keep them fed and working, we will start to see performance deterioration compared to the native crossfire and sli implementations (although lucid can always modify their design to keep pace).
Incidentally, the two lower-end hydra chips will sport a x8 connection to the controller and 2 x8 connections to the cards, and a x16 connection to the controller and two x16 connections to the cards (strictly 2x16, not configurable in any other arrangement)
The distribution engine as it is called is responsible for reading the information passed from the game or application to DirectX before it gets to the NVIDIA or AMD drivers.
So presumably it will work only in Windows, and only with DirectX games (e.g. not with OpenGL.) I'm guessing that supporting OpenGL would require a big programming effort so we won't see it soon if at all. I suspect there aren't many OpenGL games out there anyway, but I don't follow such things.
It'll be interesting to see how much extra latency the chip adds to the rendering process. I don't imagine the hardcore gamers would be too happy about it if they sacrifice an extra 50 ms to gain some FPS.
Great, now I can have 2 buggy display drivers installed at the same time, each with their own quirks. And who helps me out when I have graphical problems in a game? Do you really think ATI or NVIDIA will give end-user support for this? What about game developer support? It is a support nightmare for all involved. No thanks. Sorry, this idea is brain-dead long before it hits the shelf.
Hey, if I could just plug a second card into my system that didn't have to match the first one, I'd be at newegg right after hitting "submit". If this gets big it will definitely sell more hardware because it will lead to more frequent upgrading and just more GPU buying. AMD and NVidia would be crazy to kill this goose.
Like they give support for their cards right now. Sure, if it breaks and it's still under warranty, they'll replace it, but they rarely fix problems with game compatibility unless a majority of their users are experiencing the issue.
If vendor support is so important for you, get a console.
A software based solution to the problem of aggregating a heterogeneous collection of parallel OpenGL command streams into one, compositing the output of several graphics cards into one image, or both, has been available for years : It's called Chromium [sourceforge.net].
Although originally designed for a networked cluster with one gpu per machine, it can conceivably be adapted to one machine with multiple GPUs. Because Chromium's software based compositing would bog down a single processor system, a natural extension wo
From the Anandtech article [anandtech.com] on the subject, it appears that multi-vendor GPU scaling has been implemented, as it was demoed with a GTX 260 and an HD 4890. The mixed-vendor implementation apparently requires Windows 7 to get the card's drivers to work properly (the article was light on details on this point), but it does work. And spare me the "M$ is teh suxorz" garbage. This is aimed at gamers, and like it or not, new games come out on Windows, not Linux, so that's where Lucid's priorities will be for the
The mixed-vendor implementation apparently requires Windows 7 to get the card's drivers to work properly (the article was light on details on this point)
Windows 7 includes a new version of the display driver model [wikipedia.org]. One of the new features: "Support multiple drivers in a multi-adapter, multi-monitor setup".
In Vista, multiple adapters can only be used with Aero and the new graphics features if they all used the same driver. (The Windows XP drivers still support this, and you can still use them in Vista.)
The Z buffer shouldn't be an issue. Just render objects on each CPU (pretend they're different scenes), and then merge the images using their Z buffers as a key. You only need to exchange the Z buffer once for the final merge. It doesn't matter that objects will be drawn that would normally be hidden under objects drawn by the other GPU, because the right pixels will be chosen during the final merge based on the Z buffers.
There are certainly questions to be answered, but the fundamental idea of rendering di
There has been some noise from the big two GPU manufacturers for something similar. I don't have time atm to search for it, but I believe the article I read was on anandtech. Basically, the idea was to create a graphics card with basic functionality like 2D processing built in, but have the actual GPU chipset be user-replaceable by using a socket instead of hard soldering it to the board, so you could just plug in a new chip, and bam! instant 3D processing upgrade, without the unnecessary expense of repla
What you're not seeing is that the PCIe card **IS** the user-replaceable GPU tech! WHY would you want/need to swap out the socketed GPU? All you're keeping by your method is the ram (which is probably slow and out of date by the time you swap GPUs) and the physical connectors, which cost roughly NOTHING. In exchange you've added a ton of connections to be loose or misconnected.
Back in the Dos days video hardware was originally a register level standard. Then the accelerator companies all invented their own solutions to line drawing, BitBlts and so on. Now in Dos each programmers used Vesa Bios calls to get into high res modes but they had to write a driver themselves for anything more complex. Windows came along and acted like a software motherboard - application programmers wrote to user mode API and the graphics card manufacturers wrote drivers to a kernel level API.
can it be done in software? (Score:2, Insightful)
The article only mentions DirectX, no word about OpenGL, so it must be not a pure hardware solution. If all it does is re-routing of D3D calls, why CPU can't do it?
Re:can it be done in software? (Score:4, Insightful)
Crippleware is a common method of rent seeking, and copyrights, patents, and plain old obfuscation may obstruct genuine improvements.
Case in point: Old mainframes deliberately given a "cripple-me" switch that only an expensive vendor provided technician is authorized to switch off.
Parent
Re: (Score:2)
Re: (Score:3, Interesting)
That would require CPU. Rendering a game at 60Hz not only requires a GPU that can render your imagery within 16.7ms, but also requires the software running on the CPU to issue its directx/openGL commands in 16.7ms.
Re: (Score:2)
Re: (Score:2, Informative)
Is it useful? (Score:5, Funny)
e om
Re: (Score:2)
Forget about performance, what about power saving? For a couple of years now we have been promised the ability to shut down a graphics card and rely just on the on-board chip for desktop use, with the card kicking in when a game is launched. No-one seems to have actually implemented it yet though.
Re: (Score:2)
I know nVidia does, and I believe ATI has similar technology. There's a weak GPU onboard the chipset, but it can then switch to the faster offboard GPU when you want the grunt (at the expense of battery life).
Heck,
Re: (Score:2)
You are correct, it does exist but only seems to be used on laptops so far. I don't know of any desktop mobos that support it.
Re: (Score:2)
Re: (Score:2)
I also have a gaming PC with a more than capable second card not being used, but which would probably allow me the small performance boost I need to keep me from upgrading just yet. I think that I'm more the target market than you are.
Re: (Score:2)
I have a AMD Neo CPU (1.6GHz) and a on-board x1250 and Fullscreen Flash works well in Debian Sid.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
I have a GeForce 9500 GT, and a dual core Athlon 5050e, yet Flash fails abysmally at playing hidef YouTube on even a single 1680x1050 screen.
Re: (Score:2)
Re: (Score:2)
Suddenly, to be released to market in 30 days (Score:5, Interesting)
Re: (Score:3, Informative)
>>Finally, we can have asynchronous GPU pairing?
I think NVIDIA has some sort of asymmetrical SLI mode available on its mobos with built-in video cards. It allows the weak built in card to help a little bit with the big video card installed in the main PCI-E slot.
IIRC, it gives a 10% boost or so to performance.
Ah, here it is...
http://www.nvidia.com/object/hybrid_sli.html [nvidia.com]
Re: (Score:3, Insightful)
Cross brand or not? I'm confused (Score:2)
From TFA (emphasis mine):
To accompany this ability to intelligently divide up the graphics workload, Lucid is offering up scaling between GPUs of any KIND within a brand (only ATI with ATI, NVIDIA with NVIDIA) and the ability to load balance GPUs based on performance and other criteria.
So what is the deal? Is it cross-brand or not?
Also, they are only planning to launch their chipset on one motherboard with one manufacturer. It all sounds like a short-lived gimmick to me.
Great, can't wait until there's a Linux driver (Score:2)
Re: (Score:2)
So you can't wait until approximately...20never? You won't see drivers for anything like this in Linux. You'll be lucky to get decent bog standard 3D drivers.
Re:Great, can't wait until there's a Linux driver (Score:5, Insightful)
Parent
Re: (Score:2, Funny)
It's a slippery slope though
1) You can't read the source code to your graphics driver.
2) ???
3) You are being herded into a gas chamber.
Re: (Score:2)
And yes, I know, there is some unstable proprietary binary blob available for my ATI card which can do 3D, but it is immoral to use that
Now that you have confessed you shall say 2 our RMS and 3 hail Linus and all will be forgiven for the GNU is merciful. Go in peace, user.
(duh)
and it is actually so slow on 2D (which to me is more important) compared to the free "radeon" driver that it's ridiculous.
Or you could get supported hardware from "the other company", or stop being anal about trivial issues nobody in his right mind cares about.
Or just get a real SVGA card which is perfectly supported with completely open drivers. I hear Tseng ET3000 are a steal these days.
Re: (Score:2)
Re:Great, can't wait until there's a Linux driver (Score:4, Insightful)
I hope you die young. Seriously. If we get world hunger solved, and peace eternal, people will start to complain about even less important stuff. People complain about things, it's part of human nature. Just because 500 people died in Africa today before I got out of bed doesn't mean I don't feel that particular idiot at work is a friggin' [censored].
You can't deny people's feelings with a rational appeal to global standards.
Parent
Re:Great, can't wait until there's a Linux driver (Score:5, Insightful)
Some people rape children. How can you possibly think shoplifting is immoral?
/me steals some stuff.
Parent
Re: (Score:2)
Cute, you are trying to pull an common didactic trick by sidestepping the issue with overexerting another. I know grade schoolers like to do it:
Kid: I don't want to do my homework, it's so stupid.
Parent: For the last time kid, do your homework!
Kid: You know that hurricane Jane killed 5000 people yesterday on the west cost? And you are upset about some measly homework? How can you, there are so many worse problems on the world.
Parent: [...]
Now who the hell modded parent up?
Re: (Score:2, Insightful)
Certainly. One is important, the other is trivial nerd rage bullshit.
Performance issue (Score:3, Insightful)
Incidentally, the two lower-end hydra chips will sport a x8 connection to the controller and 2 x8 connections to the cards, and a x16 connection to the controller and two x16 connections to the cards (strictly 2x16, not configurable in any other arrangement)
DirectX/Windows only (Score:2)
The distribution engine as it is called is responsible for reading the information passed from the game or application to DirectX before it gets to the NVIDIA or AMD drivers.
So presumably it will work only in Windows, and only with DirectX games (e.g. not with OpenGL.) I'm guessing that supporting OpenGL would require a big programming effort so we won't see it soon if at all. I suspect there aren't many OpenGL games out there anyway, but I don't follow such things.
Unless the OS market changes drastically,
Latency (Score:2)
It'll be interesting to see how much extra latency the chip adds to the rendering process. I don't imagine the hardcore gamers would be too happy about it if they sacrifice an extra 50 ms to gain some FPS.
2 display drivers??? Come on... (Score:2, Insightful)
Great, now I can have 2 buggy display drivers installed at the same time, each with their own quirks. And who helps me out when I have graphical problems in a game? Do you really think ATI or NVIDIA will give end-user support for this? What about game developer support? It is a support nightmare for all involved. No thanks. Sorry, this idea is brain-dead long before it hits the shelf.
Re: (Score:2)
Re: (Score:2)
If vendor support is so important for you, get a console.
wait and see (Score:2)
last time i heard of
such magical product
it was april fool
Hardware based ripoff of Chromium (Score:2)
Although originally designed for a networked cluster with one gpu per machine, it can conceivably be adapted to one machine with multiple GPUs. Because Chromium's software based compositing would bog down a single processor system, a natural extension wo
Hybrid Crossfire? (Score:2)
Would this technology enable me to use the onboard IGP (Radeon HD3300) which
is now doing absolutely nothing as I am using a separate Radeon HD3850.
Re: (Score:2, Insightful)
The cards of a brand share drivers across a few generations. So if this solution communicates with the drivers, you get the picture.
Re: (Score:2)
Re: (Score:2)
Windows 7 includes a new version of the display driver model [wikipedia.org]. One of the new features: "Support multiple drivers in a multi-adapter, multi-monitor setup".
In Vista, multiple adapters can only be used with Aero and the new graphics features if they all used the same driver. (The Windows XP drivers still support this, and you can still use them in Vista.)
Re:fuck you all (Score:4, Funny)
I know I speak anonymously when I say that this Anonymous Coward does not speak for the rest of us Anonymous Cowards.
So, parent poster, gargle our collective balls.
Parent
Re: (Score:2)
The Z buffer shouldn't be an issue. Just render objects on each CPU (pretend they're different scenes), and then merge the images using their Z buffers as a key. You only need to exchange the Z buffer once for the final merge. It doesn't matter that objects will be drawn that would normally be hidden under objects drawn by the other GPU, because the right pixels will be chosen during the final merge based on the Z buffers.
There are certainly questions to be answered, but the fundamental idea of rendering di
Re: (Score:2)
Re: (Score:2)
What you're not seeing is that the PCIe card **IS** the user-replaceable GPU tech! WHY would you want/need to swap out the socketed GPU? All you're keeping by your method is the ram (which is probably slow and out of date by the time you swap GPUs) and the physical connectors, which cost roughly NOTHING. In exchange you've added a ton of connections to be loose or misconnected.
Re: (Score:2, Insightful)
Back in the Dos days video hardware was originally a register level standard. Then the accelerator companies all invented their own solutions to line drawing, BitBlts and so on. Now in Dos each programmers used Vesa Bios calls to get into high res modes but they had to write a driver themselves for anything more complex. Windows came along and acted like a software motherboard - application programmers wrote to user mode API and the graphics card manufacturers wrote drivers to a kernel level API.
At this poi
Re: (Score:2)
Yes? No? Yes? No!?
Dude it's about parallel processing, not quantum computers.