Printable View
Intel got no interrest in nVidia. Why would they buy them? For the soon non exsistant chipset business? For the GPU that they got Larabee for?Quote:
Originally Posted by ceevee
Plus not even to talk about how much money they had to pay. Whats your source, theinq or fuad? :rofl:
And it would be impossible to merge nVidia into Intel. Intel is the thight hardcore corporate discipline. nVidia is the free open anarchistic culture.
To start with perhaps but as heterogeneous multicore becomes the dominant architecture, it (Fusion, Cell & Nehalem) will become a threat to the discrete market and possibly the desktop market, consider that laptops now represent 53% of the PC market and that the same was said about more simplistic floating point co-processors, memory controllers and even multicore processors.Quote:
Originally Posted by Shintai
To say that CUDA has nothing to do with this is naive, nVidias progress with CUDA is all about increasing the dependence of the PC user on discrete graphics solutions in order to slow the adoption of Intel's new non-SLi compatible processor and maintain their relevance through the third round of the core wars.
It was written by this man in a private forum: http://www.seabreezepartners.net/team.aspQuote:
Originally Posted by Shintai
Trust me when I say he knows what he is talking about and would not post this for no reason. ;)
Ofcourse it will in some 10 years or so. I dont think its a secret that the ultimate goal is the system on a chip. No expansion cards.Quote:
Originally Posted by initialised
But fusion and value nehalems will be no near discrete performance the next 5 years.
Its simply not the answer to CUDA and on the long run CUDA is dead anyway. Nomatter what nVidia or anyone else want. Plain simple, system on a chip, nothing else.
I dont buy the rumourmill. There is 1million reasons to why Intel should not buy nVidia. Is there even 1 to why they should?Quote:
Originally Posted by ceevee
In short, ask yourself. What would Intel with nVidia and why would they pay 11billion plus 20-25% for it. Would it be worth it? How many years would it take to regain? Even with nVidia now doing their super profit runs we talk 10-15 years.
People said Intel would buy Nvidia 6 years ago.. It made more sense then, and didn't happen.
Now it makes no sense, and definitely won't happen.
World of Warcraft has a population of 10 million exceeding most major cities. Steam is replacing retail so yes, retail sales of PC games are way down but digital distribution more than makes up for this. Besides, the home computer has always fed ideas, innovation and technology to the console market. I don't see this changing unless and until a device with the ease of use of a games console, the application support and flexibility of a PC comes along. When MACs can game as well as PCs or when a Playstation or Xbox support PC style applications out of the box then its time to get worried about PC gaming.Quote:
Originally Posted by voltaire2030
Yes it does, what do you want, the best CPU array available or SLi graphics because you will have to compromise one to have the other by 2009 unless they bury the hatchet or merge.Quote:
Originally Posted by bowman
That's a little simplistic CUDA or similar will still be needed for coding the stream arrays on HMPUs, unless intel win with a graphics array based on specialised x86 cores.Quote:
Originally Posted by Shintai
Which begs the question: Why don't we build a city, put them all in it, and build a wall around it?Quote:
Originally Posted by initialised
Not that im saying there is any credit to his claim, but you don't always buy a company to get their technology. Remember when Nvidia bought 3dfx just to dismantle them and remove their biggest competition?Quote:
Originally Posted by Shintai
Is that a serious question? If Intel purchased Nvidia, they would get world class GPU tech instantly. They would be able to ship out Intel only systems, graphics and all, in all market sectors. And they would finally get proper integrated graphicsQuote:
Originally Posted by Shintai
But I highly doubt such a purchase would ever be approved. As for Larrabee, even the most optimistic out there are not expecting it to be in the same league as what AMD and Nvidia have now. Not only that, by the time Larrabee hits the market, AMD and Nvidia will have must faster parts out anyway.
Why thank you sir, I have a new sig.:clap:Quote:
Originally Posted by Yukon Trooper
Nvidia + Intel seems unlikely with how hostile both are right now, not to mention no financial reason to do so.
They are competing on Physics APIs, on Physics processing and on processing as a whole right now. Not that it would actually pass through any competition bureau in the first place. You can't merge a market leader with a market leader, not unless one fell through the ground.
Perkam
SLi/CF usage is so tiny its stupid. And it will always be so.Quote:
Originally Posted by initialised
You dont shell out 15billion$ against their discrete GPUs only. Just to boost AMD aswell. IGP and chipsets are already dead. nVidia cant make any for Nehalem and future.Quote:
Originally Posted by Zaskar
And that world class GPU tech accounts for how much again? Also Intel can and could make alot better IGPs if they wanted to. But its a matter of what its targetted to. Larabee dont need to compete with a 280GTX or even a 4870X2 to win. Also Intel got something none of the others have. The ability to manufactor it inhouse and optimize both design and process to one another.Quote:
Originally Posted by eleeter
So again, why buy nVidia for something around 15billion? Even all nVidias current profit and marketshare we talk a payback rate in the year 2020 or 2025. But since nVidias highly profitable chipset business is toast. We perhaps talk AMD style payback rates with anything from year 2050 till the end of the universe.
Discrete GPUs dont have a future either. And merging nVidia and Intel would be like mixing oil and water.
I feel like ATI is gonna be the king of the hill this year with a 4870X2. Yay, ATI!!!
(Now, let's start the game of pointing fingers at me and calling me an ATI fanboy... 3... 2... 1... GO!!!!!!!)
from the looks of it ATi will probably be dominating nVidia this round so the point is insignificant.Quote:
Originally Posted by initialised
GT200 is a nice chip but yields are too low for it to be a viable factor in this market and if 4870 x2 is implemented as rumored then GT200 GX2 won't even be worth considering yet alone two of them.
I'm not liking this environment, it's the kind of feeling you got when Conroe launched and you were rooting for X2 :(
Hopefully Nvidia comes back with 9800GTX+ for $199, and GTX 260 for $349.
Perkam
:poke: ATI Fanboy!!1111!!!!11 :p:Quote:
Originally Posted by Bo_Fox
9800GTX+ at $199 is not going to stop ATI from dominating the $250-$350 segment with its much better performing 4870 card, both 512MB and 1GB versions.
GTX260 and GTX280 are not going to stop ATI from enjoying its surplus sales of 4870X2 cards that perform even better than GTX260 in most benchmarks. Until Nvidia does those GT200 cards on 55nm, production run will be limited and Nvidia wont be able to cut the prices as they will be too rare, so ATI is basically winning this round. I would not be surprised if the 4870X2 beats GTX280 in most benchmarks, since the 9800GX2 is already giving a GTX280 a hard time.
Nvidia might be forced to do a 9900GTX (G92 core) with 1GB of memory and maybe 512-bit bus OR GDDR4/GDDR5 to save its plummeting profits this year! And I'm pretty sure Nvidia might be doing a 9800GX2+ on 55nm with maybe 2GB of faster memory. Otherwise, until a 55nm GTX280, Nvidia loses out badly...
Tomorrow, let's see the 4870 reviews!
I love this argument, because it is possibly the worst one ever. Yea okay, Intel could make a better IGP but they don't want to. The want to have the slowest part out there, after all that is the ultimate goal.Quote:
Originally Posted by Shintai
They will be lucky if it competes with 5 year old GPU tech in performance.Quote:
Larabee dont need to compete with a 280GTX or even a 4870X2 to win.
AMD could do the same thing, but currently it is more cost effective to mint their stuff at TSMC. Intel may be able to get an advantage here, but having a process advantage is useless if your design is much slower than your competitor.Quote:
Also Intel got something none of the others have. The ability to manufactor it inhouse and optimize both design and process to one another.
You can't make that prediction, no one knows the future to make that kind of statement.Quote:
Discrete GPUs dont have a future either.
ahh but reverse hyperthreading is real and it dramatically improves single threaded performance and k9l is just 3 months away... :ROTF::ROTF::ROTF:Quote:
Originally Posted by perkam
nVidia had their 1.5years on the top, and they treated us the consumer poorly and didn't really advance much of anything.
now it's ATi's turn
Stop rubbing it in :stick: :p:Quote:
Originally Posted by xlink
Perkam
40-45% marketshare with frequent resales. Sorry there bud. Thats the whole concept of the game.Quote:
Originally Posted by eleeter
5 year old, lol. I´m sure you saw Larabee in action, or rather not. It wont be a *GTX killer. But it dont need to.
AMD cant do the same thing, they cant even make a 45nm CPU yet. And they sold all the gear in one factory. its an empty shell.
Everything points to SoC, both from Intel and AMD.
hey I had a opteron x2 myself... admittedly I sold it at a loss a few weeks before conroe was released, and then bought most of what constitutes my current system for less than what I sold the opteron system for, but hey I was AMD all the way up until the end of 2005Quote:
Originally Posted by perkam
what is SoC though? system on chip?