There have been "hacked drivers" in the past that enabled SLI on Intel motherboards, but unfortunately they tended to lag well behind Nvidia driver releases, and, being unofficial, there were always possible stability issues.
Will intel ever support SLI?
- Thread starter slapnutz
- Start date
Actually I expect them to fully support SLI after they buy nVidia.
The nForce 200 has a few patented features which seem to be designed to compensate for low FSB bandwidth.
While the AMD 790FX has Xpress something which looks like each 16 lanes have a fast path between each other i.e x8 link #0 and #1 have a fast interconnect, so do #2 and #3, unfortunatly I forget where I saw that
.
While the AMD 790FX has Xpress something which looks like each 16 lanes have a fast path between each other i.e x8 link #0 and #1 have a fast interconnect, so do #2 and #3, unfortunatly I forget where I saw that
.
That's pretty much what I've been told by other ATi guys when asking about this. It's a matter of the benefits of doing it being outweighed by the added cost of certifying nV mobos as well, and that they'll keep on doing it for partners(like in the case of the HP Blackbird).
Silent_Buddha
Legend
Which begs the questions.
If people have been able to hack Nvidia drivers to work on Intel boards, why hasn't anyone been able to hack ATI drivers to work on Nforce boards. Or at least if they had, I don't remember hearing about it.
I'd never willingly use a consumer class Nvidia based board (although I love how well things work on the Sun Ultra 40 M2, which is completely opposite my experience with consumer boards) but I'd imagine some driver hacker would find the challenge interesting in and of itself. Even if the popularity of CF is currently much lower than SLI.
Regards,
SB
If people have been able to hack Nvidia drivers to work on Intel boards, why hasn't anyone been able to hack ATI drivers to work on Nforce boards. Or at least if they had, I don't remember hearing about it.
I'd never willingly use a consumer class Nvidia based board (although I love how well things work on the Sun Ultra 40 M2, which is completely opposite my experience with consumer boards) but I'd imagine some driver hacker would find the challenge interesting in and of itself. Even if the popularity of CF is currently much lower than SLI.
Regards,
SB
Considering that CSI was developed by Intel, I'd say they ALWAYS had the right to deny it to anyone.
The strategic question for Intel is whether a CSI license in exchange for SLI is really worth it. Or do they just want to hurt NVIDIA and drive down their price so that they can eventually buy them out...
DK
Yeah, up to a certain extend; given the size of the chipset market and the traditional dynamics there, I'm not sure not licensing it to anyone at all no matter the conditions would fly that well in court... The last thing Intel needs are more antitrust issues. However this doesn't force them to license it to NV; just licensing it to SiS might get rid of that problem. I don't now enough about these laws to be certain about any of this though, obviously.
Uhm, can I kill you for trying to put a new spin on that incredibly ridiculous rumour?
NV will not be bought be anyone and will not buy anyone else unless they can finance it with cash they already have, and not debt or stock deals. Is that really so hard to understand? :| (Damn you, Doug Freedman!)Uhm, can I kill you for trying to put a new spin on that incredibly ridiculous rumour?
I generally agree, but I learned a long time ago never say never. It certainly be very very difficult for Intel to buy NVIDIA for a variety of reasons. It is conceivable that if NVIDIA had a high stock price and really wanted to do a deal they could use stock.
Just look at Microsoft's very surprising Yahoo situation.
What's surprising about it?
Well, yeah, that's just so much against their current philosophy that I'd be quite surprised if it happened. Jen-Hsun clearly thinks he can make this a $10B company without acquiring a single profit-generating company, and Marvin always felt to me like someone who really cares about dilution. Of course, now that Marvin is retiring, who knows if his successor will share that same focus, but I'd be surprised if they went with a CFO that didn't share the same ideals and strategy...
Well nothing in that it is probably something that both companies need and there are real synergies.
Everything in that Microsoft has never done a MEGA acquisition, has never taken on debt, and has never done a stock deal. Apparently the world is coming to an end.
I don't think that is just Marv really caring about dilution rather than the company being manged from a long-term standpoint. That was the case before Marv came on board. JHH and the board are managing with a long-term philosophy, so naturally the other executives they hire will also have these attributes. IF there is a case where from a long-term standpoint it made sense to do a stock deal I'm sure they would do it.
Compare NVIDIA to say a Broadcom, which was acquiring everything under the sun. Somehow Broadcom has been able to achieve decent financial performance even though a lot of acquisitions are difficult to integrate. Broadcom used to have higher margins than NVIDIA, but over time their operating expenses have become bloated and NVIDIA is now one of the most profitable semiconductor companies. It has also always had one of the highest revenue to employee ratios in the tech industry - the highest for a hardware company that I know of - so there is a huge productivity dilution if they acquire rather than build their own. And with NVIDIA's focus on visual computing this is only going to increase. It is only a short matter of time before computer vision hits the mainstream IMO, and thats when things are going to erupt.
Actually I'd think the fact that some SLI boards overclock the hell out of the the PCI-E bus (as much as 125mhz) as well as stock gpu overclocking to this level isn't something ATI videocards would likely handle. Comes down to benchmark scores vs. frying hardware.
Proof of this being shown where, exactly....?
This is incorrect as I already stated HP's blackbird (which uses a Nvidia 680i motherboard) supports Crossfire.
It's called Linkboost and has been removed from current nForce boards.
The 9600 GT uses the PCIe frequency as it's base clock so if the PCIe frequency has been increased with Linkboost or by overclocking, expect an appropiate increase in core clock. However all previous cards, to my knowledge, had no relationship between the PCIe and core clocks and anyway Linkboost is only activated if the board detects a Linkboost capable card.
We need a tongue-in-cheek smiley.
I am aware of all of the above. It's also not as widespread as some people believe after reading "the sky is falling, evil is afoot in the kingdom" stuff with regard to the 9600GT and the PCIE clock. ATi's official stance about CF on SLi boards is that they won't do it as a general solution due to the costs and difficulties of certification, but will continue to do it for their partners, if they're asked to. Do you see any reason to disregard that?If so, why?Yes that doesn't surprise me, you're not exactly new around here are you and looking back at that thread you have posted in it.
I see no reason to disregard that but I would like to add that they shouldn't need to verify every board. Board/chipset is built to x specification, done. Simple as that. I was also sure I'd heard that Crossfire was open and that NVIDIA had simply refused to allow it.
But the mistake I made was to think you were replying to the first sentence in that quote when I should have realised you were responding to the second. Maybe in the future if you happen to be quoting the post directly above you you could only quote the particular section you are responding to
.Now may this thread continue to discuss nForce supporting or not supporting Crossfire regardless of the fact that it is quite clearly titled "Will intel ever support SLI?"
That is untrue, at least as far as ATi knows and says. And we could assume that this is something they'd make some noise about, as it would strengthen their position as the good-guys being hunted. There are far fewer illegal tackles going on than web-drama would make us believe...that does not mean they don't exist in one form or another, but they're hardly as ubiquitous as one would come to think after browsing one forum or another.
As for Intel supporting SLi...if they get enough leverage with nV they might force them to open up SLi to Intel chipsets, but that should be a helluva lot of leverage. Maybe Nehalem and its QPI will prove sufficient.
