idiotekniQues
Platinum Member
- Jan 4, 2007
- 2,572
- 0
- 76
uhhh what did ATI give away exactly?
Crossfire doesnt require anything special on the motherboard like a chip. they didnt have to give any technology away.
taltamir
Lifer
- Mar 21, 2004
- 13,576
- 6
- 76
Also tests showed that this extra chip runs very hot and consumes lots of power. And costs 30$
What is going to happen is that there are going to be the regular boards supporting CF... and then boards that support CF AND SLI for more money, power and heat...
Pretty much... the thing though, with the 4870x2, the only reason to get PCIe bridge on the mobo is if you want 4x GPUs...
I do not even want 2x GPUs... especially with the severe CPU limitation that currently exists and the inability of nehalem to overclock.
VirtualLarry
No Lifer
- Aug 25, 2001
- 56,587
- 10,225
- 126
Ok, what about the technical specifics. Doesn't the NV200 bridge chip take so many PCI-E lanes and split them into more PCI-E lanes? Does it use PCI-E 2.0 for the upstream link? What about the two downstream links? I mean, isn't the X58 PCI-E 2.0 x16 + x16 native? Will the NV200 limit this to PCI-E 1.0 x16 + x16?
And will crossfire be disabled, if you choose a mobo with the NV200 bridge chip (because it's wired into the PCI-E x16 slots)? IOW, will these mobos support both crossfire and SLI, or will they be limited to one or the other? Or will some mobos sport four PCI-E slots, two for crossfire (wired into the X58), and two for SLI (wired into the NV200). I wouldn't be surprised to see something like that from ASRock.
All of this because NVidia is greedy, and the hardware could work just fine on the native X58 chipset, if NV's drivers allowed it.
And will crossfire be disabled, if you choose a mobo with the NV200 bridge chip (because it's wired into the PCI-E x16 slots)? IOW, will these mobos support both crossfire and SLI, or will they be limited to one or the other? Or will some mobos sport four PCI-E slots, two for crossfire (wired into the X58), and two for SLI (wired into the NV200). I wouldn't be surprised to see something like that from ASRock.
All of this because NVidia is greedy, and the hardware could work just fine on the native X58 chipset, if NV's drivers allowed it.
Toonces
Golden Member
- Feb 5, 2000
- 1,690
- 0
- 76
Any proof to your two statements above?
If you are privy to the actual details of the AMD v. Intel CrossFire licensing deals please share them with the class, it would be interesting to see how much AMD values CrossFire (or conversely, how little Intel values it).
SunnyD
Belgian Waffler
Speculation: ATI (Pre-AMD) gave Intel a CrossFire license in exchange for the GTL+ FSB license - which was why and how ATI used to make Intel chipsets.
More Speculation: Assuming the agreement is still in place, there's nothing saying AMD/ATI can't still make Intel chipsets. Which is part of the reason why after the AMD/ATI acquisition, Intel so hastily moved towards QPI and away from GTL+. As you can see with the Core series, GTL+ still has a lot of life in it for everything except the highest-end server parts.
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
Nobody is going to buy a M/B with the NV sli chip on it . Why should they . When M/B makers can just add these little babies.
http://www.engadget.com/2008/0...gpus-for-powerful-mat/
http://www.engadget.com/2008/0...gpus-for-powerful-mat/
Oh dear...that's quite a poor name choice considering AMD already picked the Hydravision name for their multidisplay features.
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
What does intel have permission to use . XF and SLI??? XF is open . SLI is not . Intel didn't get permission to do anything . NV is going to market a chip to the M/B makers to add their chip to the mix . NV has NO QPI.
If you can't see the usefullness of this chip your blind . Because thats what this chip is. Its Blind it doesn't care what GPU U are it just allows all chips to scale . not just ati XF or NV sli. If It scales as they say . Your statement about it being dead befor its born is FALSE.
If it delivers as advertized . This Intel backed . X intel engineers in Isreal using an old intel fab has a great future.
http://www.lucidlogix.com/news/pr_2008_07_14.html
If you can't see the usefullness of this chip your blind . Because thats what this chip is. Its Blind it doesn't care what GPU U are it just allows all chips to scale . not just ati XF or NV sli. If It scales as they say . Your statement about it being dead befor its born is FALSE.
If it delivers as advertized . This Intel backed . X intel engineers in Isreal using an old intel fab has a great future.
http://www.lucidlogix.com/news/pr_2008_07_14.html
Keysplayr
Elite Member
- Jan 16, 2003
- 21,219
- 55
- 91
What makes you think the things you do?
AFAIK, there is still a partial ring bus being used. They changed the arch to hub.
Definitely seems to be an improved memory controller, superior to the ring bus.
And I need to ask you why the hype for this lucidlogix Hydra chip when you have never seen it in action? What are the limitations if any? What are the bugs?
Do you know?
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
Extelleron
Platinum Member
Posts: 2753
Joined: 12/26/2005
Good that we are finally going to see SLI on an Intel chipset, and seemingly nVidia is giving up on building Intel motherboards, finally. Once Intel chipsets have SLI... the only reason to buy an nVidia board is gone. Unless of course SLI on X58 is going to be crippled compared to what it would be on an nVidia board. In that case it's just dirty dealing by nVidia as usual.
It still pisses me off that they are forcing Intel to use the power hungry nForce 200 chipset for no reason whatsoever. You don't see a special ATI "Crossfire chip" on Intel boards to make Crossfire work. There is no need for such a chip other than cluttering up the motherboard and increasing power consumption.
Whos forcing who here? Intel didn't give NV QPI now and I am betting not ever. NV has nothing Intel wants. Intel got nothing from NV . NV got nothing from intel. Other than A sign to put around NVs CEo neck saying firesale going out of business in the not to far future.
Inorder for SLI to work with x58 NV has to sell this other chip . The M/B makers will have higher cost because of the board layout and running traces.
NV forced intel to use SLI chip on Skulltrail that doesn't work for shit.
This time intel forced NV to use Intels chipset if they want to sell to intel buyers using intel chipsets. Lets take bets here. I say we never see an intel branded M/B with the sli chip.
Intel can simply have the hyper chip installed ( Intel has put money into this company, people and product)and wait for NV to block it. Which would spell disaster for NV sooner rather than later If the hyper chip only scaled 50% thats better than sli scaling . Now if sli was infact faster I don't know how it would turn out . But if the hyper chip is better faster and performs as intended. NV would be forced out of business by the consumers..
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
Yes it does keys . Amd seems to have improved on the ring bus. Which brings up the fact that larrabee is using a ring bus. I wonder how intels version will compare to ATIs. I am taken bets.
Sorry keys missed the second question. Iam not hyping it. Its new , its exciting and its kinda ontopic .
Now I answered your question . Now you ans mine .
Why are you hyping SLi on the X58 when we know skultrail sli using said chip sucks bad . Try running sli with 2x 280's and see what happens. So are you in the habit of pushing products we know suck based on past performance. If your talking about another chip a newer one . When you have never seen it in action? What are the limitations if any? What are the bugs?
Do you know?[/quote]
SunnyD
Belgian Waffler
While I read this, the only thing I could think this would be good for is a couple words in it's description - distributed computing. Ergo, the GPGPU tasks that have nothing to do with graphics may be possible with this. I am really curious about their claims of linear and above-linear scaling, especially with graphics tasks, since SLI and CF both can't even achieve linear scaling. Further, I fail to see how an additional piece of logic being a sort of arbiter can even hope to achieve linear scaling performance if it's playing "middle-man", incurring it's own additional latencies, etc.
Imagine if it does work however, and it allows multiple GPUs to be used in a distributed manner, what then? What about processing delays? What about mixed performance? How will they alleviate "input lag" and "microstutter" and all that? The only way I could see is if they use the lowest common denominator in the mix, and take into account their additional latencies. I seriously... SERIOUSLY doubt we'll ever see linear scaling, let alone better than maybe a 20% gain.
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
Well you won't have all that long to learn more abot hydra as its going to be at IDF.
INTEL DEVELOPMENT FORUM I wonder why its at IDF????
Sunny this is from the link . I think it pretty much says what its going to be used for among other things.
The real-time distributed processing engine, called HYDRA?, offers a groundbreaking approach to scaling 3D graphics performance in a multi-GPU environment. Outlined for the first time on the new Lucid Web site at www.lucidlogix.com, the novel technology is the industry's first to work with any GPU, any CPU or chipset and on any application that gamers and professionals may choose. See above its at IDF . So I guess its pretty safe to say these guys get QPI.
INTEL DEVELOPMENT FORUM I wonder why its at IDF????
Sunny this is from the link . I think it pretty much says what its going to be used for among other things.
The real-time distributed processing engine, called HYDRA?, offers a groundbreaking approach to scaling 3D graphics performance in a multi-GPU environment. Outlined for the first time on the new Lucid Web site at www.lucidlogix.com, the novel technology is the industry's first to work with any GPU, any CPU or chipset and on any application that gamers and professionals may choose. See above its at IDF . So I guess its pretty safe to say these guys get QPI.
SunnyD
Belgian Waffler
Well I understand all that. But still, the claims of linear scaling... wow. Of course that statement is ambiguous enough to mean a linear scale of 1% per card rather than 100% per card. I guess we'll see.
[/quote]
I can't be bothered to go and re-read the Skulltrail stuff on AT, but didn't Skulltrail mostly suck in general?
How much of SLI performance is due to the use of the chip, and how much is due to the generally unsuitable platform?
krnmastersgt
Platinum Member
- Jan 10, 2008
- 2,873
- 0
- 0
I wouldn't mind the HYDRA engine actually doing what it says but how many people honestly think it'll achieve linear scaling? What makes them think they can do it when the people that made the cards themselves couldn't? And Nemesis, you might want to tone your rebuttals down a bit. In the end everything can look good on paper, like a lot of people have said already, let's wait for it to come out before we judge it to be a failed idea.
Nemesis 1
Lifer
- Dec 30, 2006
- 11,366
- 2
- 0
I agree with Ya . We don't know how its going to scale with Hydra. I am just thinking in tech terms that my glass is half full. Nornally I am pretty open to all hardware advances.
But at the same exact time. I get pretty pissed . When people say that X58 gets sli. When infact the x58 gets a sli chip . We all know how that turned out on skulltrail . Hay I was excited skulltrail got sli till I seen its performance:thumbsdown:
Than you have a thread like this . That says X58 gets sli' Well thats great when you see the glass is half full. I like that. What I don't like is when Something like Hydra is introduced into the thread . The thread still memains on topic. But the halffull glasses that cheered sli on x58. View the hydra glass as being half empty. Just doesn't ring right.
Either your for advancing tech . Or your not. Hydra at this early stage looks to be something we should all cheer for. One for all all for one. Thats how this chip works . It doesn't care if it's NV /ATI or intel gpus . Now to me thats progress and its refreshing.
If it works almost all would agree with my point of view. Would you not?
But at the same exact time. I get pretty pissed . When people say that X58 gets sli. When infact the x58 gets a sli chip . We all know how that turned out on skulltrail . Hay I was excited skulltrail got sli till I seen its performance:thumbsdown:
Than you have a thread like this . That says X58 gets sli' Well thats great when you see the glass is half full. I like that. What I don't like is when Something like Hydra is introduced into the thread . The thread still memains on topic. But the halffull glasses that cheered sli on x58. View the hydra glass as being half empty. Just doesn't ring right.
Either your for advancing tech . Or your not. Hydra at this early stage looks to be something we should all cheer for. One for all all for one. Thats how this chip works . It doesn't care if it's NV /ATI or intel gpus . Now to me thats progress and its refreshing.
If it works almost all would agree with my point of view. Would you not?
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter