This is already the big bang HERE.
http://en.expreview.com/img/2009/10/...RINERGY_02.jpg
This is already the big bang HERE.
http://en.expreview.com/img/2009/10/...RINERGY_02.jpg
http://www.guru3d.com/news/msi-big-b...board-delayed/
few months left only
Lucid is more than likely competing in the "Worst company and FUDspreader in the IT industry 2009" competition :D
Perhaps they should announce a final delay, and a release date somewhere in 2012 :D
its weird seeing radeon and geforce in the same computer. never thought something like this would materialize.Quote:
Originally Posted by NapalmV5
i dont have a problem with hydra but i am still skeptical about it. do we have real numbers other than ones from hydra? dont you have to have win7? does it support d3d11, opencl and cuda? it might have the same problems xfire and sli did at their beginnings. just looking at how different ATi and nvidia's gpu's are. who knows if it will be perfect. will the object rendered on the other gpu look different than the other? they do use different algorithms. it would be cooler if they could take the advantages of both architectures and render the parts that they work best on like one super gpu.
You mean like this puppy here?Quote:
Originally Posted by Chumbucket843
http://www.planetdognine.com/feature...iles/bfast.jpg
it was never advertised as better than linear, thats impossible. It was advertised as "near linear performance" is the direct quote.Quote:
Originally Posted by Hu1kamania
better than linear would be the card getting 110% performance with hydra, which is obviously impossible.
I remembered reading better than linear from a couple difference sources.Quote:
Originally Posted by Decami
From FUD:
Not that I am counting on "above-linear" performance or anything, just that it was once a claim. I find it pretty far fetched, but not impossible. Even linear scaling across 4 GPUs w/full PCIe bandwidth would be magnificent.Quote:
That's some big claims for what so far has been an unknown company, and it gets better. Not only does Lucid claim that Hydra is "delivering another industry breakthrough: near-linear to above-linear performance with two, three or more graphics cards," but that this can either be done by integrating the Hydra SoC either onto any motherboard or to an add-on card
i believe that would be an act of god.Quote:
Originally Posted by Hu1kamania
this tech sounds awesome, but the delays are stating to make me worry. i can't imagine what might have caused a 3 month delay one week before launch, when they announced the launch date EVERYTHING should have been ready. also, why launch the chip on a p55 board? wouldn't an x58 make more sence as x58 owners are more likely to have a multi-gpu set up?
+1Quote:
Originally Posted by 570091D
Dream on. So you believe some small company can do something MUCH better in a few months than two big companies in ~3 years? (+ add 3dfx experience before that!) :ROTF:
To be realistic, it will offer:
1) mixing cards, worse scaling than SLI / CF;
2) proper PCI Express bandwidth for P55 multi card cofigs.
That's it. All benefits over traditional SLI / CF.
One can dream ;)Quote:
Originally Posted by 570091D
Linear gains does not mean 100% scaling. It just means that adding another graphics card will always add the same amount of performance. Above linear means that adding another card and another card will add more performance than the previous card.
Linear performance gain for each card: x -> 1.5x -> 2x
Above linear: x -> 1.5x -> 2.2x - > 3.5x
neverending 30 days lol
Delayed till November afaik. Or even longer...Quote:
Originally Posted by eric66
Umm, yes it does unless you've just redefined the word linear.Quote:
Originally Posted by munim
In any case it is possible to have super-linear scaling if the addition of the second card results in a reduction of overhead in the entire system. That obviously isn't happening with today's technology though.
If you chart a linear relationship between two variables and then connect the data values, you will have a straight line. This does not necessarily mean that as one variable increases by one unit, that the other variable also increases by one unit. It simply means that as one variable is increased by some amount, that the other variable also increases by some amount, every time. Don't even begin to argue this. Google = friend.
linear is this:
first value is x
second is x+y
third is x+2y
chart them and you will have a strait line. 100% scaling is not needed
in th real world if scaling isn't 100% then it would probably be a logarithmic regression. its not like the overhead of communication would be linear. load balancing would only get worse with the more gpu's you add. thats why you see 70-80% faster with 2 gpu's and 3 gpu's is ~130% faster in a game that scales well. this chip does not use magic as far as i know.
you know, its funny. I have been following lucid since they announced it exists, reading most reviews and direct information from lucid interviews themselves. The only time I have ever seen someone say "above Linear" is in taht single fudzilla news post. hmmmQuote:
Originally Posted by Hu1kamania
I actually remember the first statement released by lucid, posted here in this very forum, and there was an argument many pages long, with a bunch of people debating on what "near linear" performance could be. never remember "above linear" though.
your entirely correct. Except for one thing, your forgetting on what your basing this on, a GPU.Quote:
If you chart a linear relationship between two variables and then connect the data values, you will have a straight line. This does not necessarily mean that as one variable increases by one unit, that the other variable also increases by one unit. It simply means that as one variable is increased by some amount, that the other variable also increases by some amount, every time. Don't even begin to argue this. Google = friend.
Actually, a system is linear if it follows the two properties of linearity:
1) scaling. if y(t) = x(t), then Ay(t) = Ax(t) | If you scale your input by A, your output will also be scaled by A.
2) superposition. if y1(t) = x1(t) and y2(t) = x2(t), then y1(t) + y2(t) = x1(t) + x2(t) | If you add two inputs, your output will be a sum of them.
Yes, linear scaling means 100% scaling.
Lol wtf? Those formulas are just a bunch of useless bs.Quote:
Originally Posted by m0da
Linear means a constant change. So if card x gets 100points in test one, and then gets 150points in test two, it will get 200points in test three. Notice that differences in points? 50point increase with each test.
So if you gain 50% in performance going from 1 card to 2, you will get 50% more performance going from 2 cards to 3. This just means constant scaling. Not 100% like some of you guys think. (Especially m0da who thinks that if he can put up some ridiculous, poorly-formatted formulas, he can dissuade people from continuing the topic in hopes that they won't understand, and accept that he is right).
If my table gains value at a rate of 5% per year, and its valued at $100 at the first year, the next year it will be $105. Then $110. Then $115. This is a linear relationship. According to you people who think its 100%, it would be $100, then $200 etc.
?Quote:
Originally Posted by Salavat23
5% per year, is gain of 5%...($100>$105>$110) <-- thats not 100% scaling, that just linear progression. :rolleyes:
123
If you do not have the educational background to comprehend these formulas and properties, please do not insult them, calling them bs. I can scan these exact formulas from a textbook, but then again, you wouldn't understand what you are seeing, hence there is no point.Quote:
Originally Posted by Salavat23
Contradictory to your point of view, linear scaling does in fact mean that scaling your number of cards, scales your output by the same amount.
See the scaling property of linearity, referenced in my earlier post.
Salavat, if you would like to learn, with proof, more of what you may think you know, feel free to send me a PM.
Let's not bicker any further.
Moreover, I'd appreciate if you would refrain from further insults to the property of linearity.
Precisely.Quote:
Originally Posted by Xoulz
Well, that's still more expensive than a SLI/Crossfire link:rofl:Quote:
Originally Posted by Chumbucket843