Why AMD should merge with Nvidia
Comment To save us from Intel
TO QUOTE A famous expression: "there can be only one". Looking at how things shape up for AMD and Nvidia, it's rather obvious to see that in the future there can be only one.
To battle Intel and its upcoming technologies, it's going to take a lot more than what AMD or Nvidia can offer on their own. They need to form a very close cooperation or even fuse into one single, across-the-board company.
Daamit is in big trouble with its CPU line and Nvidia is in big trouble with... pretty much everything. But AMD has a big debt coming along with it and Graphzilla is far from being the richest boy in the classroom. However, they're both green, so check one on the compatibility list.
Technology match
Not only in colour are the two companies a good match. Now that Ruiz is gone,
maybe a bit of common ground can be found. If only Huang can forget about his
ego for a few moments, only enough to realize that a common future is a much
better path than a separate future. Daamit has a very nice engineering
potential as well as several fabs in its pocket. It has the most efficient GPU
to date and a CPU line to go with it, embattled as it may be. Nvidia on the
other hand, can fill in the high performance GPU gaps and can offer additional
chipset know-how. Together, they have what it takes to deliver strong single and
multi GPU technologies, a very complete physics support for gaming purposes and
a great GPGPU range. AMD already delivered Puma, a unified mobile platform.
Together with Nvidia, they can put the pressure on Intel, big time. They got
Fusion coming up, Nvidia has Physx and if they work together that can mean a
world of pain for Chipzilla.
A blue future
Arguably, Intel has all the aces right now. It's going to be a blue future for
the IT world, but also a blue mood for Envydia and Daamit if Intel can make
things work in reality the way they currently work on paper. And, unfortunately
for the greenies, in the past few years Intel has actually made things work even
better in reality than what they promised in slide shows. The upcoming discrete
GPU from Intel, Larrabee as well as the upcoming Nehalem assault should give
Damvidia a lot to think about.
For now, there is no reason to think that Intel's new aspirations in the GPU market will be a fluke. Of course, there is no reason to think otherwise either. Larrabee is an unknown factor right now. But Intel has proven many a times that even though it might be a fresh entrant into a given arena, it can make up for the lost experience extremely fast. This is a very agile and versatile competitor, definitely not to be underestimated. Huang knows that and that's the reason he's so passionate about Intel lately.
Under the belt
What's even more worrying is the fact that Intel will undoubtedly hit both green
camps under the belt with Larrabee. For now, they appear to accept the fact that
rasterization is still hip in the gaming world, but Larrabee has the potential
to do much more, they made sure we all know that. First of all, there's
raytracing. This may be the ultimate in providing super realistic graphics and
fantastic special effects. and Intel is banging the war drum on this subject.
Nvidia and AMD have nothing in this area as of right now and they will pretty
much wake up not only being beat at their own game, but even falling behind with
a few years from a technological point of view.
To this, we must add Intel's know-how in the field of chip design. Larrabee is already supposed to be multi-core, and very powerful too. Think about the raw amount of power offered by the beast as well as the efficiency with which that power will be used. These things alone should be more than enough reason to worry for the greenies.
Conclusion
It's been a while since AMD has had its own GPU line. The latest exponent of
this line, the 4800 series, is one of the best pieces of graphical acceleration
engineering to grace the IT world in the past few years. But one cannot argue
that Nvidia is a very strong ally to have on board. Even with its lack of
charisma, Intel's Santa Clara neighbour is what AMD needs to weather the storm
ahead.
It goes the other way around too. With a mounting debt and a host of other problems, Chimpzilla can't afford any more months filled with bad news and migration of key personnel. A fusion or at least a very close partnership is something that will undo some frowns and give the investors reasons to cheer about. It would boost motivation in both camps and bring numerous marketing, logistical and technological assets under one hopefully happy roof. Nvidia, with its monstrous GPU tactic is safe, because DAAMIT's targets in this arena are different anyway. So why not put two plus two together and give Intel some real competition in the future, not two small companies which to wipe the floor with? µ
Comments
Do you have a brain ?
This is the most idiotic and stupid article of all times. Do you think when you're writing, Intel's Larrybarry is just a joke in your face !larrabee ray tracing
http://www.pcper.com/article.php?aid=532Atleast Carmacks says larrabee ray tracing is a cold turkey for few years to come. Doesn't exactly congratulate that quake ray trace-demo..
Not possible...
Dude, you must be living in another world to believe in such a story. This is so impossible... anyway have a nice time while dreaming of merging your fav companies :)What are you smoking?
Are you serious? There is absolutely no incentive for AMD if they merge with nVidia. nVidia is slipping really fast, but I agree that nVidia needs AMD.I really don't see the point of saying that Larabee has so much potential because GPGPU also has the same potential and it is being realized "NOW". For all we know, Larabee might suck or it could be the other way around. So please don't tell us that the future is bleak for AMD and nVidia because of Larabee.
There is no doubt that Intel is on top of its game, but that doesn't mean that AMD has inferior technology. It is more like, AMD has average tech while Intel has very good tech. The same goes for nVidia. Now, if you would, only for a brief moment, put yourself in the real world. There are only a handful of people who demand the very latest and greatest CPU for the desktop, majority still needs the most ordinary CPU for what they do. The money is in the mainstream. The same applies to GPUs.
For what it's worth, nobody knows for a fact if Larabee will whoop a**. However, I can tell you right now that AMD's and nVidia's GPU's are very good compared to a non-existent Larabee.
Cheers!
I dont see it
AMD's graphics cards will vastly outperform Nvidias stock this year by having higher performing/cheaper cards at the price points that matter. So that brings to question why bother with Nvidia? Nvidia is a sinking ship, and I can't see anything in your article apart from maybe physics that would imply otherwise.AMD would be better merging with VIA for their ultra-low power tech rather than nVidia if you ask me!
What?
And give the resulting abomination a complete monopoly of the graphics market?Shurely not.
Lets Debunk.
Ray-tracing:Ati/Nvidia GPU already do it. Cinema Industry is doing it since the days of the 2900XT.Larrabee: 32 cores P54C X86 SSE 4 Arquitecture. Even at 45nm, it will be warm, slow, and never forget, to fully use it, it will have a proprietary API. Chrome, PowerVR, GLide Already showed us that you should free the API. DirectX still up, but they are all dead.
i740: For those who don't remember. It is not the first time Intel tries to enter vga/gpu world.
Moore's Law: Intel Larrabee might as well fail this one. Dont forget GPUs ussually have a 6 month life/developing cycle. Im curious about this also.
Conclusion: With luck, Larrabee will be a bit more than a paper launch. My regards to the author that spilled names and terms and talked about merging. Dont worry mate, in 10 years we will still have Nvidia, DAAMIT, and Intel in the game.
Intel can't make a working driver
Let me know when Intel releases something involving graphics and it comes with a working driver actually capable of doing all they promise.They have failed on everything from i740 to all integrated chipsets. Their CPU's are AWESOME currently, but their GPU's have sucked forever (mostly due to drivers). Unfortunately for Intel, AMD/Nvidia seem to have bought all the good driver writers already...LOL.
When your track record is 100% failure, chipset after chipset (and i740 on top for discreet) I think they don't have much to worry about on the green side yet. On paper larabee looks pretty good. But it takes a driver team to make that design work, and to date, Intel has failed miserably there.
That would make things worse
AMD is still trying to get over the problems that the merger with ATI created. When you are going down, you need to lose weight, not gain even more. Besides there are very few things that Nvidia will offer to AMD in terms of GPU technology.Via
Nvidia should gulp down Via. The they would have their future tied to either X86 maker. IBM should gulp down AMD. Heck they could make a Cell/X86 hybrid that would eat the universe.uhhh
Ever heard of antitrust? The two single largest discrete GPU manufacturers (in their market they don't have other competitors) merging? How on flying fuck earth do you expect that to happen?Not only is it very unenticing, I like the idea of three-way high-end competition, it's impossible. Please stop posting wishes and dreams as editorials, I like this site otherwise.
Drugs much?
You called it spot on with the AMD + ATi merger, but this time around I think you guys should keep the new guys away from writing sensational articles like this - amusing never the less. April 1st was a good while back tho :PI would question your more than recreational use of drugs and other mind altering substances.
Change of underwear?
Not sure where you have been living until now, but i remember what happened last time Intel raved on and on about how great the graphics card was going to be and we all know how that turned out.Before we start reaching for the change of underwear it might be worth seeing Intels Grail in the flesh first.
Can't wait for your next article on crossing a dog with a cat.
Smoke
I wish I could get some of the stuff this guy is smoking. :)Why not ?
actually it's pretty well thought.AMD is in debt...
The CPUs are mobing a little better, but not as good as they should
Nvidia is kinda cornered with its only sword, the GPUs
They are not going to make chipsets for Nehalem
They're in the mobile market, but no one knows if they're gonna hold
And intel rules the market.
The Core architecture just blew the whole IT world away and they're making billions while AMD and Nvidia are just staying afloat. But Intel is like an iceberg and soon the tip (Larabee) might scratch the competition.
we'll have to wait and see
nonsense
Utter nonsense. There is no reason at all for them to merge. You say AMD has fabs, nvidia has PhysX? WTF? The fabs are used for CPU's, not GPU's and PhysX is hardly a reason to get into bed with nvidia.A better deal would be for TSMC to buy nvidia.
If all else fails...
it may be a reasonable thing for these two to combine forces, at least from the standpoint of the enemy of my enemy is my friend or at least my ally. Nvidia for all its bizarre engineering decisions regarding heat dissipation has been good at making video drivers. While ATI for all it's new found GPU goodness has been exceedingly vile at getting decent drivers out the door. Maybe between the two entities someone might actually get worthwhile Vista drivers out the door.The most ridiculous thing EVER
You should go to work on wall street or for gartner. This is the most idiotic thing I have read in a long time. Intel has bombed at everything other than CPUs FOREVER. They are 50% overstaffed, with lots of people who say NO.Even if Laffabee works, who cares. Every time one of you moronic "press" people pronounce that something new product or technology in the future is gonna be apocalyptic, it never is, and is most often a dud.
Just absurd.
And by the way, ever heard of Hart Scott Rodino..... I thought not.
This will never happen
... for one reason: AMD hates Nvidia and Nvidia hates AMD. They are direct competitors, they could never merge anyway.A good deal would be...
A good deal would be... McDonalds buying Nvidia and start making burgers on those gigantic hot Nvidia chips... :-pAre you on drugs man? Havent you check the news lately ? 4870X2 kills the 280GTX... so why would AMD/ATI need Nvidia??? What for??? For making burgers? LOL
Name one...
people seem to forget the absolute hash already that intel has made of their graphics chipsets (& especially the drivers) in the past... and so all of a sudden they are goin to get it right this time ???aye blimmin right... and mebbie Via will introduce a quad core killer !!!
Not gonna happen
When AMD and Nvidia has both their own GPU lines, I don't think that they would get that much benefit from merging their GPU lines. Nvidia-VIA sounded more interesting, some epia boosted with hdmi nvidia could be a killer.Good...
Lets merge the designers of unstable CPUs (AMD Phenom B2) and unstable motherboards (Anything NVidia) and pit them against the designers of dead stable super clockicing higher performance CPUs and dead stable motherboards...AMD/NVidia will have the fastest platform the 0.5% of the time it works...
Guys listen up!
Read the SIGGRAPH paper on Larabee and understand it and then compare it to the current GPU architectures and you will see why Larabee is the biggest scary thing for the green arrogant goblin. Because architecture wise it is far superior if they can pull it off NV is in big shot of trouble besides their G84&G86 fiasco and their deadend SLI road. And people in the industry know it read Tim Sweeney's comment and even I don't have a clue Carmack doesn't diss the design because he knows it is awesome and a major evolutionary step ahead of all that GPGPU bollocks currently out there.Looks like fun...
But why does DAAMIT need nVidia anyway? I know you've mentioned a lot of technologies merge which should benefit both of them. But that would almost be impossible to happen. DAAMIT doesn't need nVidia's CUDA though since they're moving to OpenCL for their GPGPU platform.I don't know anything about laws. But isn't it has to be DAAMIT to acquire nVidia and not the other way around because of that x86 agreement with intel? So, dude, where's the money?
I think what DAAMIT needs now is push its "Torrenza" into the market. Because that's the area that hasn't been dominated by intel. And if it's got any luck, it could even surpass any platform offered by intel right now in term of computational power. And focus more on mobile market which has the highest growth rate lately.
Oh, and BTW, I doubt Huang will buy your idea.
AMD next CPU...
AMD needs nVidia ?I don't think so.
Chipset: 780G, 790GX, 790FX, SB750.
GPU: HD 4850, HD 4870, HD 4870-X2
CPU: Phenom B2/B3... well... damn it!
AMD next CPU...
Phenom (65nm) 2 Mb L3:
Idle = 39.4W
Load = 104.1W
Deneb (45nm) 6 Mb L3:
Idle = 32.4W
Load = 57.3W
Smaller, cooler, faster, cheaper.
Preview: http://tinyurl.com/62phf5
Slow or...
Not? Okay, clear up some details shall we, ATi has 800 arithmatic logic units, each capable of doing two mathematical ops per clock, they are clocked at 750MHz.Nvidia has 240 of which are capable of doing 3 ops per clock, and are clocked at 1.2GHz.
Larrabee has 32 cores - each of which have a 16 ALU wide vector unit. 32x16=512 ALUs capable of 2-4 ops per clock. Larrabee is predicted to be running at somewhere around 2GHz. Ladies, do not underestimate it.
It's possible
There is a lot of poorly informed naysaying here. Larrabee is looking really good - talk to some of the ISVs who are working with it. The development team is NOT the same as the team working on the integrated parts. If you talk to them, you'll find these people really know what they are doing, and all indications so far are that it is going to be an impressive part.In contrast to the commentators here, I'm sure AMD and Nvidia are very worried about what is coming, and well they should be. I can certainly see the possibility of a merger happening a few years down the road.
Can there be a net gain?
On the face of it, AMD/ATI + Nvidia merger to battle Intel may work but would it in the real world?The potentials: Nvidia's engineer can help speed up development of products, the unification of SLI/Crossfire, CUDA/whatever AMD is going to do, Physx and Hovak.
The bottom line, where is AMD or Nvidia going to get the money? AMD is under 3B market cap and Nvidia is around 6B! Would be possible for Nvidia to buy AMD but why would AMD want to hop on a sinking ship now (even ignoring the x86 licensing issue)?
Nvidia may barely have enough cash reserve to deal with its short terms problems: http://www.theinquirer.net/gb/inquirer/news/2008/07/31/hp-pays-half-nvidia-problems
I suppose AMD can try to scoop up Nvidia after that blows over? :D Nvidia could become pretty cheap assuming Nvidia stock price continues to plunge and the lose a ton of cash...
Nonsense Intended!
Haha! I love the sense of humor at the INQ, and from the comments, I see the readers are on par!We're talking billions of dollars in equity here, mate! You don't go selling, buying or merging billions of dollars on top of a maybe-possibly-good-despite-terrible-history future product! Even being from Chipzilla!
Users have pointed out the anti-trust obstacles already, and suggested the more reasonable NVIDIA+VIA, TSMC+NVIDIA and DAMMIT+IBM transactions. Those would be more synergetic/complimentary solutions, while DAMMIT+NVIDIA would be very redundant and monopolistic. What holds DAMMIT's perspectives up is the graphics division. If it needs reinforcing, it's in the other businesses, so NVIDIA wouldn't make much sense. Also, NVIDIA already has business involving VIA and ARM that could spark some love stories.
DAMMIT cannot cede fabs for NVIDIA, as it's struggling financial status forces them to spin those off or sell them away: the so-called Asset-Light policy we've been hearing from Ruiz for the past year. TSMC/UMC, on the other hand, have the foundries and the first already produces the chips for both GPU camps. The idea of a TSMC+NVIDIA merger producing ATI chips actually brings us to the bottomline of all this complex situation: "The enemy of my enemy is my enemy!", contrary to a previous comment.
Up until now, the three players (DAMMIT, NVIDIA and Intel) have been competing and collaborating on different fronts among themselves. Intel + Crossfire, NVIDIA's chipsets, Intel's Skulltrail with SLI, etc. Yet, you have the duels: Intel vs. AMD, NVIDIA vs. ATI, and in a commercial sense, NVIDIA's SLI vs. Intel's QPI licenses drama. The soap opera goes on and on.
As we progress into the GPGPU, Larrabee, Havok, Firestream scenario, it appears things are beginning to deteriorate. NVIDIA is dropping out of the chipset business, Intel is stepping into the GPU/GPGPU industry, AMD is taking the OpenCL path, etc. I expect things to get rough among them, and despite us consumers getting cheaper innovative products as fruit, it's the possible fall of any one of the big three and the consequent stagnation of the industry that worries me. AMD almost fell, causing those Intel EE (Expensive Edition) processors reach the US$1,500+ mark. The weakening ATI during the HD2000 and HD3000 moments made NVIDIA believe it could sell the GTXs at the US$600+ level! I guess that close call shows the picture.
But following the rationale behind the article that sparked such yadda yaddas, we could say Apple should buy Linux to defend itself from Windows 8 (which could be a good product given it might not be based on Vista).
Nonsense intended!
bad bad idea
the overruling idea in a merger should be synergy. There is no synergy in this, both produce gpu's and chipset, by merging you would need to kill off one set of them. Worse it would not solve the problem, the combined company would not be more competitive than now.amd in itself is not in a bad shape, it cannot compete simple by $, intel has the money to develop a lot op projects, if one fails (prescott etc) they have options (core)
plus intel can afford to build and modernise it fabs way faster than amd. If amd had sufficient cash to develop and build the long gap between the last athlons and new phenoms would have been shorter and 45 nm would have been in place. Ams needs a cash rich buyer that is willing to slog it out and if possible has fabs. IBM looks like a logical, one of the korean big boys another.
nvidia and via would be a nice match, it would still be a fabless operation but, kill the via chipsets and build something that will kill atom/945 combo and make money that way.
Regulatory Approval Fail
AMD/Nvidia would never get regulatory approval for a merger until there is another strong player in the GPU market. Intel is a few years away from being that player, and their success in the GPU area is certainly not a foregone conclusion.SuperLarrabbe?
No,no,no...that's a very ambitious thought,but in no way even close to what's going to happen.AMD/ATI is currently better than nVidia,first due to the 4800 series and secondly due to nV being a sinking ship(arguable-they must have smth up their sleeves!).From any point of view,you can't know that Larrabee will be a hit-but even if it is-both ATI and nV will deff be able to answer with something of their own,if they don't have smth already out that's better or competitive with Larrabee.What I (want to) see in the near future is 3-way competition.And that's better than 2-way!What?
Intel should stick to CPUs. Focus on the core buisness, and leave the vid chips to the people who can both create and support them properly. ( Not that Nvidia's been doing that very well lately either ) Intel makes good CPUs, sometimes better then others, but all around a consistantly decent product.As far as AMD and Nvidia, I don't know how many of you guys/gals have used the Quadro and FireGL cards back to back, but quality wise, ATI is far superior. Nvidia is a little faster for rendering, but ATI's picture quality rules for design. Not to mention their cooler, and quieter for the most part. ATI just needs to market better, as always. What is it with AMD companies and marketing anyway?
What's interesting will be AMDs Fusion and how it may fit into this conversation. CPUs that can offload physics to a GPU cell. A GPU cell that can offload raytracing to the CPUs. This will be the future, and AMD/Intel are already hasing out the boundries. Nvidia is more agile then we're giving them credit for, but they are in for some serious downsizing. AMD doesn't have to do a thing but have a good product, and they will gain from Nvidia's regroup. If Nvidia sticks it out, Intels graphics division will see some serious competition from both camps since neither Intel or AMD can do anything with Nvidia.
After that, you never know. Maybe via will take over the Laptop space all together with Nvidia graphics :)
@JDocs - What motherboard has Intel EVER made that was stable? (Ok, so there are a few but) I have a large pile of them right here, and the tech tickets, from 1993 to today that would prove you wrong. I'll admit that there is a decent one every once in a while, but no more then any other company. Asus has actually been our most reliable motherboard vender, and I have one of their boards on the bench right now that caught fire. Tells ya something.
Would Need to be Bigger
It would take AMD/ATI, NVIDIA, IBM, and VIA to merge to have a shot to dethrone IntelName
And what name should the "fusion" have? DAMNVIDIA?nvda need to go on offensive
NV is facing competitive pressure, and potential financial trouble. The biggest problem is that Intel has figured out the gpu game, and coming after NV. So there is no point to play ball with MSFT and their directx stuff.A better way to come out of this is to launch preemptive strike and open up the GPU - basically to capture developer mindshare before Larrabee could.
I 100% agree
Days gone by, when the world was competitive for intel and they were fighting off AMD and trying to produce thier ill fated 740i... I would agree with the BILE SPITING detractors is see posting.But the posters must be young in the years and quite fanboi-ish of their red and green boxing idols. Because if you look at Intels history... They learn. Then they research, then they stomp.
Same here now.
Amd kicked Intel's teeth in with the integrated memory controller and more efficient IPC. But Intel did what Intel will do again. They waited. they learned, they researched and then destroyed their competition. hell they did it without using a on-board controller.... they did it with what AMD and nvidia does not have... billions and billions of R&D, and multiple teams working in house .. separately in competition with each other to build the best.
Now we see in q3, that Intel will release Nehalem... that will top their own products by 30% per clock... and allow for almost 30% more clock speed to add to that.
This CPGPU business is the same. Intel has watched people build Phyisx gpus, gpus, gpu-api's for parallel processing ... and waited.
smart move.
Now that intel knows what works, and what doesnt... (in this case all that shitty custom api crap to use "GPU power for other tasks") and is now ready to release thier x86 compatible parallel processor.
Skoff if you will. but if you do you are ignorant. you should be just as afraid of intel as Nvidia is. And moving your target product like Amd is. is it coincedence that amd has gone Nasa on us? making many chips for cheaper and slapping multiple chips on a card and making them for the mid range crowd? No, Amd knows whats about to happen. And is in survival mode.
Not you Daddies intel kiddies....
This is not the same intel that made the craptasic p4. this is the extremely efficient design Intel that is kicking the crap out of amd , and in q3 will put then in the dog house on a choker. This is the intel that is smartly producing a easy to code for x86 graphics adapter with massive vector units... and a FULLY PROGRAMMABLE CORE.
Something that amd and nvidia are YEARS FROM HAVING.
No, it will not be almighty in the 1st generation.
But Gen 2 will be.
There can be only one
This applies to everything when you think of it. There can be only one... pencil. That's right, there is only one brand of pencil! There can be only one... asphalt company. That's right, there is only one asphalt company! There can be only one... liquid oxygen supplier. That's right, there is only one company that supplies liquid oxygen!What a maroon.
Dumbest
A Dammit/NVDA merger is pure Hollywood, wresting Jim Carey of Dumb and Dumber fame to secure the title of "dumbest". C U @ Oscars!Yep
Why don't we just get it over with and start calling it DAAMNvidiIT?Pfft!
There's nothing I hate more than clueless engineers venturing into armchair MBA. Not to mention that all we need is yet another merger, like it's any good for customers.where are you drashek ??
jeeeeeeeeeeeeeeeeez wtf ?? sorry axonn but drashek (on a bad day !!) would make more sense than this article does, yikes O_O
Last Resort...
greetings to all you caveman's living under a rock.The only possible solution is to have a three way competitive market for cpu/gpu. Ideally, it would be for nvidia to merge with IBM solutions this will create a new powerhouse in the market which will consist of amd/ati, intel/centrino, ibm/nvidia. This will only help the consumer in the end which is what everyone should be going after, remember the more options we have the merrier.
cheers chaps
YEA!!! What are you smoking?
So how would a merger like this help the consumer?It would not…
ATI & Nvidia are already in trouble about fixing prices, and now you want to kill competition with a merger?
I would hope the SEC comes in and takes control of things
It'll be a flopper.
Larrabee is destined to fail. The similarities between it and Itanium are horrendous. Stop and consider that by the time Larrabbe arrives it will be late, slower than expected, and nVidia and ATi will have cards 2-3x as fast as now.Intel cannot make a Core2 style rise in the GPU arena, it is too competitive and too fast moving.GPGPU is the future.
Big visions
An Nvidia article at Inq without total bashing!? I'm sure Charlie would never have approved this.Anyway, I don't see how AMD's and Nvidia's technologies could be merged. At this point, it would create antitrust cases too so it's not possible.
Are you on pills?
ATI has demonstrated that 4000 series likes raytracing too. Where the heck did you pulled that out... ?.."Intel will do raytracing and AMD/nVIDIA cannot do anything about that" It was demonstrated on Computex that 4850 did full raytracing and managed to score 60 fps at higher resolutions than Intel did with Larrabee.I hoped you will treat every news with objectivity... but as far I can see... you are turning blue...
Stop acting as journalists...and start being you... circus boys.
its not time yet
Anyone rember the nvidia/3dFX deal?thats whats going to happen.
Mergers, mergers everywhere!
The author of this article and some of the commenters are buffoons. Lots of people seem to have a misperception that merging two multi-billion dollar companies is easy. One commenter wrote:"Nvidia should gulp down Via. Then they would have their future tied to either X86 maker. IBM should gulp down AMD. Heck they could make a Cell/X86 hybrid that would eat the universe."
Eat the universe... yeah ... right.
If one thing that recent history should teach these idiots it is that merging two companies, even those that have good synergy, is HARD F'ING WORK and is far from guaranteed success. Just look at AMD's takeover of ATI or Daimer's takeover of Chrysler, Time/Warner of AOL and the list goes on and on. AMD spends something like $100M/Quarter servicing debt from the ATI acquisition. That money could be going to shoring up its microprocessor operations. Instead, that money is going to investment banks who brokered the deal. THat is not good for AMD and not good for shareholders (are there any left?).
You people who are advocating mergers willy-nilly should think about this stuff before running your mouths (keyboards, whatever). Merging companies has lots of challenges, least of all egos and technology pissing contests that go on between the merger and mergee.
I'm just glad that most of the people making these idiotic comments are probably surfing porn and commenting on The Inquirer from their parent's basements where they can't do any real harm.
Laugh!
I have one word for the team working on this.. Pentium 4.Look at that utter failure of a processor and you expect that design team to work on Larrabee? Are you serious?
I smell another Intel Failboat.
maybe
if Larrabee comes out with working drivers (ha Intel, with working graphics drivers on a whole new paradigm HAHAHHAHAHAAHAHAHAHAHAHAH)...and at 32 cores (if that many) in 2009-10 say 18 months from now. That's like 3 generations in GPUville...
multiply the ops/sec all you want, doesn't make it real world. 4870x2 * 2 in quadfire = could beat it a year before it exists "in theory" That's w/o taking into consideration that crossfire/sli doesn't scale perfect... why would larrabee? I know my quad core machine doesn't use the extra cores most of the time. On GPU, texture goes unused in math heavy games, and math units go unused in heavy texture games.... In a non-highly optomised game I'll bet larabee = FAIL.
At best larrabee might be competitive, otherwise it'll be equiv to Via/SIS/Matrox/Intel SUPRISE! I'll stay skeptical til I see it benched.
also... big question, what's the price of it going to be?
It made sense for AMD
and nVidia to form a strategic partnership before the ATI buyout. Senor Ruinz was apparently more interested in playing Monopoly with real money. If it had worked well he would be one of America's top CEO's but instead he's on his way out to pasture. Do AMD fanbois stick pins in El Hector dolls at night?As for the notion "there can be only one" it's conventional wisdom in economics that the optimal number of competitors is three which would imply that Intel should be welcomed to the party and nV fangyrls should say a prayer before bed for their fave GPU vendor.
Larrabee sounds very interesting for GPGPU since coding should be more general and hopefully Intel will sponsor a solid open source toolset to motivate adoption.
Shanghai doesn't seem to be generating much excitement or Charlie would write an article a day about how good it's gonna be. If AMD doesn't produce at 45nm then the lifeboat might get a little crowded. Someone could buy AMD for pennies on the incredible shrinking USD and spin off ATI when Nasdaq is next on a roll.
let's hope not
While I do hate the big whit and blue bastard that is Intel, if AMD and Nvidia's coming together did just so happen to stomp out Intel, there's much of a chance of EVERYTHING going up in price.With AMD being the only CPU provider (excluding VIA and SUN), and there being no competition left in the GPU market, you can very well expect the price of your build to sky rocket.
And what happens if the two feud over profit? Then we're back where we started. Thus giving Intel a chance to re-build it's empire.
Not the brightest idea
There is no reason whatsoever why Nvidia would want to merge with AMD. Nvidia has always been the leader in everything graphics. Proof is their technology, market share, new product pipeline (upcoming mobile graphics.. major wins to be announced soon, physics based products, CUDA ... the list goes on). Nvidia's balance sheet is pristine with $1.5 Billion of cash and no debt. True Nvidia has hit a bump in the road with overheating GPUs but this does not mean that it needs to merge with an almost bankrupt company like AMD. AMD and ATI could have a great product lineup now it does not make them a serious player when they are facing both Intel and Nvidia with deep pockets and good technology. As I see it, it is too late for AMD to get itself out of its problems and Nvidia should focus on its strenghts rather than take additional weaknesses of a major merger with AMD. Nvidia should stay out of the PC CPU market too, The PC CPU market is history. PCs are in their last gasp. Once developing countries like China and India finish buying the PCs they need there won't be any demand left for PC CPUs. It is this acknowledgment by Intel that got them to come up with a mobile device called ATOM and a their first ever GPU (in two years from now) called Larabee. Intel moving in on Nvidia's term is an acknowledgment of Nvidia's vision and technology leadership. I vote for Nvidia to keep doing what is has always done and that is to create the product that will bring about a better visual computing world. Were we left with Intel we would still be running with CPUs with higher megahertz.Cassandra
Briefly, Cassandra is a foreteller from Greek Era - Troy whose predictions eventually came true, but cursed so no one would believe her.Back to the article. I agree with the author on the merger. May not happen this year or next, but if and once Larabee becomes as successful as Core archi, then things will really make a U-Turn change.
Lets just say intel has the resources (research manpower, cash and technology) to make Larabee a Bona Fide threat to both green and red camps.
Read the signs. Red is cash-strapped but has the CPU and GPU tech. Green has cash, but only has GPUtech. Seems to me like one is compliments of another. Like the case of 3dfx,
only time will tell.
Are you dumb?
One word for you: Antitrust. Merger will create a monopoly in the discrete graphics market. SEC will never allow it.Not to mention that any change of control in AMD will jeopardize the x86 cross licensing with INTEL, which means that NVIDIA CEO have to give his spot to AMD's CEO, are you kidding me? Jun-Sun is known to be one of the biggest egos in the silicon valley. What did you smoke before writing this? Wait a second, sorry, I forgot, it's THE INQUIRER, not the most serious site around... sounds more like bar chatting between drunks.
Oh dear
Saving us from Intel by merging with Nvidia and ATi is a but like stopping the forest fires in California by burning the whole place to the ground in advance.Ati doesn't need to merge with Nvidia because Nvidia is busy flushing itself due to the soaring costs of its product failure rate.
They also have a downward spiralling share price, and a product set which is neatly out priced and outgunned by ATi.
Personally I would rather be eaten by Intel than let Nvidia survive.
Nvidia should go and merge with oblivion and gives us all a break.
interesting..
..very interesting reading. but would it be allowed? could be considered a graphics monopoly...Spot On
Short term (6-18 months) I don't see it. But later out, I think this is a spot on analysis. The major roadblock is not the logic in the forecast but the arrogance and pride of the heads of Nvidia and AMD/ATI and the fear and greed of the stockholders. Plus how bad the economy gets and for how long. That will certainly get the two companies closer together if not merged if they continue to screw up royally in a downward economy that will punish both companies more harshly than normal if they continue to foul up. It may come down to (which is the whole point of this excellent article) a matter of survival.Sad but true
I started reading the comments and realized it is a very long list indeed. My oh my, the entire readerbase has rung up and called this heresy. If they can't feel superior due to their choice of GPU what else is there to feel superior about. Still, as among other things a 3d artist, being able to do raytracing on something other than the cpu is a major breakthrough indeed and damvid can't ingnore this. I don't know if a merger is good or bad but it does seem likely that intel is leveraging its chipmaking might here. I believe if intel really wants to, it can do this. And whaddoyouknow, they might even a write a driver for it, god forbid.ITS SACKING TIME!
Charlie if you read this could you have this guy fired he is an idiot or nvidiot i should say, and he does no justice to the The Inquirer. What a shitul article. The regulatory issues alone make such a prospect impossible. And where is it in AMD/ATI's interest to merge with FAGVIDIA!Not going to happen
We all know that the only real need for a stand alone GPU with it's own memory is for playing 3D computer games, otherwise an integrated chipset is fine for work such as document editing, e-mail, and surfing. The one thing no one has mentioned is that games manufacturers support two types of GPU's- NV and ATI. Few games will actually run on Intel Chipsets as it is. Games manufacturers aren't motivated to support some intel GPU that Intel claims is the Bees Knees. Intel is great at one thing- CPU's. Unless they buy NV, they won't be producing any GPU's that will work anywhere near as well as ATI or NV and there is no logic to NV and ATI merging to battle some pathetic Intel GPU. Keep on smoking that droe buddy. Pass it around. LOL!!!To ALL IDIOT who thing NV/ATI already have raytracing.
To ALL IDIOT who thing NV/ATI already have raytracing.Search Google by this key: "Ray Tracing Wiki" & "Phong and Gouraud shading wiki". Look at those pages at wikipedia.
Now, do you really think NV/ATI could do that images in ONE-PASS shot at 60 fps??? How MUCH IDIOT do you have to be!!!!!
If you don't know how much computational power required to to make ray-tracing, phong & gouraud shading, then you certainly doesn't know what you're talking about.
As to IDIOT who says ray-tracing is not important. Why do you think all high-end (AUTOCAD is NOT a high end product) for CADAM/CAE/CAM use OpenGL??? Do you really want to board an airplane that was design using Direct-X and Raster shading?
Do you really know when you play those oh so great game, when you walk by a tree those tree doesn't change perspective as what you do in a REAL WORLD.
What good is PhysX computation when you don't CORRECTLY drawn EVERY photon ray (that is LIGHT if you don't know it) in EVERY pixel on that screen??? Mind you than wood reflect light differently than steel, stone, asphalt, etc. And that's must be calculated in EVERY light reflection/refraction on those pixel has!!!!
Do you really think to drawn a REALISTIC image so it RIVAL the REAL PHOTOGRAPHIC IMAGE can be done in just 1, 2, 4, 8 GPU card??? Why do you think ILS, when making TOY-STORY movie, it took them 2 YEARS just to render a 90 minutes movie with their HUNDREDS of PC's??? And still... IT DOESN"T RIVAL THE REAL WORLD IMAGE.
To those who REALLY knows what RAY-TRACING is all about they WOULD KNOW what LARRABEE mean for them.
AMD knows it. NV knows it. That is why Huang is smokin like he is. And AMD scramble their Organizational units.
Just small application of Ray-Tracing:
- You go to a car dealer. You choose your option for your car and the dealer render it on the monitor so it would shows you just HOW IT WOULD LOOK LIKE EXACTLY WHEN IT'S ATTACHED on your vehicle. Including all the TRUE COLOR under the SUN, STREET NIGHT and whatever. So you in ADVANCE knows whether your selected BLING-BLING would looks right on your car BEFORE it is even delivered to you!!!!
- You design wardrobe, you ran Ray-Tracing and found out EXACTLY how those wardrobe would look like. Whether the placement of the stage lighting would make it BETTER or WORSE beforehand. Whether you should use WHITE colour lamp or a NATURAL colour lamp. Whether the wardrobe reflection would make it better or worse in stage show/real use.
You and I could think a million and more for the application of Ray-Tracing. It's the way forward. If you look at MINORITY REPORT movie, do you think all of the advertising boards showing live ads could be done without Ray-Tracing? All the Tom Cruise Advance-3D GUI screen drawn without calculating Ray-Tracing? Give me a break....
However knowing it's benefit, do mind this. It would maybe take yearssss.... before GPU could do this. I WOULDN'T HOLD MY BREATH for it.
Only when they solved how to put hundreds/thousand processor in 1 die. Until they solved how to move trillion of data within those hundred and thousand gpu cell in/out in nano/femto second.
Don't expect a REALTIME PHOTO-REALLISTIC IMAGE GPU.
TFLOPS calculation require data move in/out in Terra Bytes also. Otherwise all those OH SO GREAT/OH SO MANY GPU cells would be wasted.
In the end, I do agree with the above article since even both company R&D spending could not match what Intel has. They need to complement each other if they want to survive beyond Larrabee. Again, Intel is a MUCH bigger company than BOTH of AMD/NV.
Although given their EGO, I don't think they would.... Unfortunately...
Funny
It's funny that less than 1 year ago people were saying how terrible AMD was and that they were a sinking ship. All companys have their ups and downs.apple buys nvidia, ibm buys amd
nvidia quits chipsets for intel/amd makes chipsets only for mac-intel/via/arm and graphics card for mac/intel/amd/via/armamd quits processors for windows and makes processors only for linux. amd makes chipsets for intel/amd/via/arm. amd makes graphics cards for intel/amd/via
this sounds more reasonable but still basically a idle speculation and baseless rumour
I dont think Intel would do well
Intel is great at making cpus. In the Pentium age, intel sucks preety bad only when AMD came out and smashed the pentium around, same performance, half the price ad lower power consumption. intel only got good when the core 2 comes out. but cpu is very slow compared to gpu because cpu works in parallel compared to gpu, who works in series which is much faster. second the ati 4850 have 800 processors compared to i think 32? processor in Larrabee. i think the biggest blunder is that intel tries to stick a pair of core 2 duo onto a graphic card. CPU SUCKS AT PROCESSING INFORMATION COMPARED TO GPU, and intel isn't good at making drivers. with Nvida's CUDA and OS that support GPU as the central process unit, CPU wouldn't be so important anymore. i think larrabee as do or die. if Intel fails this badly, Intel might go down. :) i think larrabee will come out in 2 years? if that is true. both ati and nvida have time to make 2-3 more generation of graphic card. its like having ati 2x 4870x2 onto one card and 2x 280gtx all in one card. i think ati and nvida shouldn't merge, but share some of the drivers like physX and Havoc(sp) and have better programs so they would operate at its besti hope intel falls so intel would stop monopolizing
Keep the spice with a dead Bee
Keep the spice with a dead BeeI'm sure you can do better than this fusing paid commercial into some article.
Even Chippy doesn't believe that its Lara Bee would even stinger properly to the competition in higly claimed gpu sector.
Lara is produced as a side effect (P54C), just like Atom, just to keep with spice flow. They have 45nm for free a year leap ahead, so it's the pay off time for them. And new 32nm won't be that free for them just like 90nm bandwagon wasn't. Hm, i wonder if this 45nm will have some serious bugs like nv 80nm chips on dopamine.
And lara should have at least 64 cores cause the shrinkage of the old .8um p54c chip certainly allows it, and then it might be worth buying it if cores would reached promised 2GHz. Anything else it's just deception for overwhelmed enthusiast to fill up some coins in Chippys pockets.
@Lightnix
you do much of multiplying ;-), maybe you should divide something Is that "16ALUs in one core" or is it "quad pumped 4ALU" from old NetBurn tech? P54C with AVX steroids is just a replica of an obsolete processor with one FPU unit replaced with ALU so it's 2 ALUs capable of 4-way SMT, and SMT NEVER scaled better than 50% still promised 2-way SMT in the ngma Nehalem (or araound 12% in old NetBurn P4) so it's 32CORESx2ALUs "x 4-way SMT percentage" x core clock. And for some reason they fake or really have an issues to make this "monster" (khm. really an ant) to work @ 2GHz