Podcast: Play in new window
| Download
There is a funky dance going on right now between chip giants Intel and nVIDIA and I just want to cut through the crap and tell you that no matter what the companies are saying it is likely to end with nVIDIA being purchased by Intel. Both parties know it and the only thing that hasn’t been determined yet is the price, which is what all this posturing is about.
Intel this week cancelled Larrabee, its proposed graphics processing unit (GPU) that was intended to compete with both nVIDIA and ATi (now a part of AMD). The moment AMD bought ATi Intel had to decide whether to build or buy its own GPU to stay in contention. They decided to build, or at least said they had. It’s hard to say how viable Larrabee ever was but at some point it turned from a weapon against nVIDIA to a barrier to Intel buying nVIDIA. So Larrabee had to go, because without that chip Intel presents a much less imposing target for the Department of Justice and Federal Trade Commission which might oppose a merger on anti-trust grounds.
With Larrabee, Intel could be seen as crushing a major rival. Without Larrabee, Intel is just trying to enter a new market.
Intel wants nVIDIA not just for its GPU’s but also for its mobile chips. Mobile is a big part of the future of computing yet Intel in 2006 cancelled its business aimed at handset makers, writing-off $5 billion in the process. Remember that? Despite attempting to back-in to the market again with its upcoming Moorestown post-Atom low-power processors, what Intel would really like are nVidia’s superior Tegra AXP chips aimed at mobile phones and media players. Ironically nVIDIA didn’t even design the Tegra processors, which it bought a couple years ago with Portal Player.
Intel had to do something the minute AMD bought ATi. Now with Larrabee gone Intel has no real choice but to buy another company to remain in contention. The only such company available is nVIDIA.
Notice how Intel has been making a serious effort lately to settle the anti-trust cases pending against it, especially with AMD, to which Intel is paying $1.25 billion. Yet against nVIDIA, the legal action actually appears to be heating-up. Why would Intel hold an olive branch for all the others yet still oppose nVIDIA? Again it is to drive down the price of an eventual acquisition and Intel has nothing to fear from an nVIDIA legal case if its actual intention is to buy the company, retiring the lawsuit.
But the action is not all coming from Intel. Last month nVIDIA announced they were suspending development of chipsets for new Intel processors. Later we read that nVIDIA was going to release a chipset for Intel’s Lynnfield processors only to have Intel question the validity of nVIDIA’s license to the Intel’s Direct Media Iinterconnect (MDI) technology, which connects Lynnfield processors to the chipset. Both companies are talking tough and so far nVIDIA has not released a product that supports Intel’s Core i7 or Core i5 processors, yet without Larrabee Intel really needs that support no matter how much they fume. Meanwhile nVIDIA has its own Ion low-power System-on-Chip for portable Internet devices and, inspired by AMD/ATi’s Fusion forthcoming CPU/GPU hybrid, nVIDIA has been working on a similar chip of its own. Intel was, too, with Larrabee, but now that’s over.
Intel could effectively block the nVIDIA hybrid processor through the MDI licensing ploy, above, and doing just that would have worked to the advantage of both Intel and AMD while Larrabee was still viable. I suspect this may have been a big part of Intel’s reason for settling the AMD anti-trust suit. It may have been part of their argument to AMD, too, about why the smaller company ought to settle at this time, because doing so would remove an nVIDIA threat. The big question is whether Intel knew even then that Larrabee was doomed? If they did, then the real plan was for Intel to absorb the nVIDIA hybrid processor and make it its own, which they can effectively do now that AMD has promised not to testify against Intel under any circumstances as part of their legal settlement.
That would be Intel turning the tables yet again on AMD, which may have been suckered.
If this reads like a huge conspiracy theory that’s because it is.
But sometimes conspiracy theories are true.
There is a funky dance going on right now between chip giants Intel and nVIDIA and I just want to cut through the crap and tell you that no matter what the companies are saying it is likely to end with nVIDIA being purchased by Intel. Both parties know it and the only thing that hasn’t been determined yet is the price, which is what all this posturing is about.
Intel this week cancelled Larrabee, its proposed graphics processing unit (GPU) that was intended to compete with both nVIDIA and ATi (now a part of AMD). The moment AMD bought ATi Intel had to decide whether to build or buy its own GPU to stay in contention. They decided to build, or at least said they had. It’s hard to say how viable Larrabee ever was but at some point it turned from a weapon against nVIDIA to a barrier to Intel buying nVIDIA. So Larrabee had to go, because without that chip Intel presents a much less imposing target for the Department of Justice and Federal Trade Commission which might oppose a merger on anti-trust grounds.
With Larrabee, Intel could be seen as crushing a major rival. Without Larrabee, Intel is just trying to enter a new market.
Intel wants nVIDIA not just for its GPU’s but also for its mobile chips. Mobile is a big part of the future of computing yet Intel in 2006 cancelled its business aimed at handset makers, writing-off $5 billion in the process. Remember that? Despite attempting to back-in to the market again with its upcoming Moorestown post-Atom low-power processors, what Intel would really like are nVidia’s superior Tegra AXP chips aimed at mobile phones and media players. Ironically nVIDIA didn’t even design the Tegra processors, which it bought a couple years ago with Portal Player.
Intel had to do something the minute AMD bought ATi. Now with Larrabee gone Intel has no real choice but to buy another company to remain in contention. The only such company available is nVIDIA.
Notice how Intel has been making a serious effort lately to settle the anti-trust cases pending against it, especially with AMD, to which Intel is paying $1.25 billion. Yet against nVIDIA, the legal action actually appears to be heating-up. Why would Intel hold an olive branch for all the others yet still oppose nVIDIA? Again it is to drive down the price of an eventual acquisition and Intel has nothing to fear from an nVIDIA legal case if its actual intention is to buy the company, retiring the lawsuit.
But the action is not all coming from Intel. Last month nVIDIA announced they were suspending development of chipsets for new Intel processors. Later we read that nVIDIA was going to release a chipset for Intel’s Lynnfield processors only to have Intel question the validity of nVIDIA’s license to the Intel’s Direct Media Iinterconnect (MDI) technology, which connects Lynnfield processors to the chipset. Both companies are talking tough and so far nVIDIA has not released a product that supports Intel’s Core i7 or Core i5 processors, yet without Larrabee Intel really needs that support no matter how much they fume.
Meanwhile nVIDIA has its own Ion low-power System-on-Chip for portable Internet devices and, inspired by AMD/ATi’s Fusion forthcoming CPU/GPU hybrid, nVIDIA has been working on a similar chip of its own. Intel reportedly was, too, with Larrabee, but now that’s over.
Intel could effectively block the nVIDIA hybrid processor through the MDI licensing ploy, above, and doing just that would have worked to the advantage of both Intel and AMD while Larrabee was still viable. I suspect this may have been a big part of Intel’s reason for settling the AMD anti-trust suit. It may have been part of their argument to AMD, too, about why the smaller company ought to settle at this time, because doing so would remove an nVIDIA threat. The big question is whether Intel knew at the time that Larrabee was doomed? If they did, then the real plan was for Intel to absorb the nVIDIA hybrid processor and make it its own, which they can effectively do now that AMD has promised not to testify against Intel under any circumstances as part of their legal settlement.
That would be Intel turning the tables on AMD, which may have been suckered.
If this reads like a huge conspiracy theory that’s because it is.
But sometimes conspiracy theories are true.
Tags: AMD, Intel, nVIDIA
This entry was posted
on Tuesday, December 8th, 2009 at 6:22 am and is filed under 2009.
You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
Yes Larrabee was an impediment to a merger, but not the primary one. I don’t see how there is any chance at all that the sale would be approved by the FTC or the EU. Intel is the #1 graphics and chipset vendor, nvidia is #2 and, AMD is #3. #1 and #2 will not be allowed to merge. nVidia will have to be bankrupt before it has a chance of getting through the regulators. I mean seriously, the EU is holding up Oracle buying Sun because of MySQL!!!
Let me add, when AMD bought ATI there was essentially ZERO overlap in their products. That is why they got through regulatory hurdles. AMD was not making IGP or standard chipsets in any significant quantity. You cannot say the same for Intel and nVidia. They are direct competitors with lots of overlap in a 3 player market.
Wonder if this is more about physics processing unit (PPU) then video/GPU cards?
nVidia has recently (or rather in the last few years) has really struck out on the way of building processors for the Mobile Market – starting with the Tegra?
The interesting thing is that Tegra is powered by Intel’s arch nemesis – ARM based CPU. Intel has divested its own ARM-based assets (selling the PXA line) to Marvell. Intel has decided clearly to pursue all markets including Consumer Electronics (CE) using Intel x86 Cores – e.g. the Intel Sodaville CE3100 is powered by the Intel Atom in addition to many other hardware features. So to add to the conspiracy theory is – it may help fight the ARM onslaught may be by killing Tegra upon acquisition of nVidia.
Just such a thought was germinating in my head when I heard of the Larrabee delay, but your explanation of the multiple plays to drive down nVidia’s share price and appease AMD really make (eve-il) sense for Intel.
Poor nVidia; you had it all to lose a few years ago, and have done a smashing job.
The most interesting question to me, if this came to pass, is what would the combined company’s open source strategy be?
Intel is the most pro-open-source of the big three GPU companies — they employ several of the core X.org developers, and their open source drivers have been the most reliable and complete for the last five years. And the same pattern holds in their other product lines. nVIDIA, on the other hand, is the most antagonistic. They provide no open source drivers, no chip specs, no nothing.
This matters to Microsoft most of all, because graphics drivers are one of the major stumbling blocks to Linux’ desktop penetration. (That’s probably why Intel funds X.org development.)
nVIDIA’s stance towards open source has more to do with their belief that they have valuable intellectual property in their driver source code than any fundamental Linux vs. Windows viewpoint.
If Intel buys nVIDIA their stance toward open driver development will most likely remain the same with their current graphical offerings. The nVIDIA stuff will most likely remain proprietary if in fact there is IP in those drivers which may benefit ATi if exposed.
NVIDIA seemingly to prefer their name to be spelt all in capitals these days – ftp://download.nvidia.com/image_kit/LG_NVCorpBadge.pdf (yeah yeah, weird but there we go).
[...] author and industry veteran Robert X. Cringely thinks Intel will buy Nvidia. With the AMD/ATI sale, increasing popularity and importancy of GPU processing, and the gap closing [...]
You have no fucking clue about what is going on in the industry. Intel and Nvidia hate each other more than they hate AMD. Not to mention AMD tried buying Nvidia before ATI and it failed miserably due to Nvidia’s CEO having unreasonable demands.
“it failed miserably due to Nvidia’s CEO having unreasonable demands”
LOL. The smarter guy wanted to run the company, that’s unreasonable? Turns out it would have been a much better move than keeping the criminal and inside-trader Hector Ruiz at the helm.
AMD failed to buy NV? do you think NV was “peanuts” like ATi?
better check market value/captial/etc before saying ” unreasonable demands”
best regards,
“Both companies are talking tough and so far nVIDIA has not released a product that supports Intel’s Core i7 or Core i5 processors, yet without Larrabee Intel really needs that support no matter how much they fume.”
I’m not convinced that’s true. Intel chipsets work fine for most purposes – SLI is a niche market and although NVIDIA chipsets generally have better integrated graphics, the buyers of integrated graphics do not appear to be very bothered by the poor performance if Intel’s integrated offerings. For the majority of buyers who want decent 3D performance but don’t need SLI an Intel chipset with an NVIDIA or AMD/ATI discrete graphics card is a perfectly good option. Intel really don’t need NVIDIA’s chipsets.
Intel has a near term roadmap to move its integrated graphics on die for the low end of the market. Without Larrabee it doesn’t have a clear path to bring high performance 3D on die but the need for that is still several years away and in the meantime discrete cards from NVIDIA and ATI will meet the needs of gamers.
This would be a great article if Larrabee were in fact cancelled.
Has Cringely ever actually been right on a prediction?
2006
• A huge expansion of .Mac to one TERABYTE per month of download capacity per user
• 2 new Intel Macs with huge plasma displays, but with keyboards and mice as options — literally big-screen TVs that just happen to be computers, too.
2007
• Apple releases iTV, a bunch of flat-panel MacTV’s that contain Mac Minis.
• The year the net crashed (in the USA). Video overwhelms the net and we all learn that the broadband ISPs have been selling us something they can’t really deliver.
2008
• Microsoft will indefinitely extend the life of Windows XP.
• Not only will Bill Gates be retiring from Microsoft in 2008, CEO Steve Ballmer will, too.
• Apple will introduce a subnotebook/tablet computer/media player
• Along the same lines look for OS X to bifurcate clearly into two lines — Mac OS X and plain OS X (for devices like the iPhone) with Apple licensing (non-Mac) OS X to a few companies, including Sony.
• Apple will build into some Macs support for the Windows API, allowing Mac and Windows apps to run side by side with no need for virtualization software except to run Linux.
2009
•The next Yahoo CEO will dismember the company and sell it piecemeal, made possible by the fact that only the Internet companies have much real cash. The Yahoo name will survive but the company will not.
Try going back farther with your research and study of Cringely predictions. The primary purpose of the predictions is to have fun and stimulate thought. If this column doesn’t inspire some thought, maybe you are in the wrong line of work.
IBM, HP, and Oracle are all in a head to head race to see who can buy what. Many years ago when Lotus bought cc:Mail, Microsoft ran out and bought Network Courier. This stuff happens in every industry all the time.
I’m still waiting for all his BS products, like the tin foil hard drive that will revolutionize the industry.
Larrabee 1 (the chip) was cancelled because it didn’t perform anywhere near the level of the competitors from NVIDIA and AMD. It has absolutely nothing to do with a potential buyout.
Tegra was primarily designed by NVIDIA, using NVIDIA’s design flow. Bits and pieces were melded in from the various companies that NVIDIA purchased over the years, but it’s NVIDIA’s baby.
Finally, the only way that an Intel/NVIDIA merger would happen is if Jensun was retired or was made CEO of the whole shooting match. Do you honestly see that happening?
Listen to dude, he’s got it right.
Dude speaks the truth – no matter how great Intel’s marketing department is, they cannot make up for Larrabee’s poor performance.
Larrabee was probably canceled because nobody wanted to buy it. Apple certainly didn’t.
The SoC (System on Chip) you mentioned that NVIDIA has, is actually called Tegra, not ION. ION is basically a 9400M graphics core and other core logic for a MB. Great read and interesting speculation.
Oops… made myself look like an idiot! That’ll learn me to comment b4 reading
Apple will buy NVIDIA that or after the EU blocks the Oracle/Sun merger then Apple will buy Sun
I know your post was meant to be funny and since Sun has nothing of value to Apple that’s not going to happen. Sun will go to some vulture investment co that will dismember it. But Apple buying NVIDA is interesting.
Apple bought PA Semi a designer of PPC and possibly ARM chips. Now Apple does use ARM in some of their products, and I’d be very surprised if the “tablet” is not ARM based. But Apple also uses NVIDA and the tablet probably depends on something from NVIDA that would be best for apple if it wasn’t on the open market. Also Apple despite using Intel chips is still a competitor to intel in the embedded world because PA Semi’s PPC chips are all over the place. Yes Apple “gave” Intel the lightwave stuff but it would be just like Steve to give a peace offering to someone before turning their world upside down. This seems far closer to reality than a one off joke.
[...] [...]
[...] Cringely thinks that Intel (INTC) will be buying nVIDIA (NVDA). Story here. [...]
[...] Intel will buy Nvidia, I, Cringely says Cringely believes that despite what Intel (INTC) and Nvidia (NVDA) are saying, Nvidia will end up being purchased by Intel. Reference Link [...]
Since this is now getting media play, let’s set something else straight. This claim:
“which they can effectively do now that AMD has promised not to testify against Intel under any circumstances as part of their legal settlement.”
is simply absurd. AMD did nothing of the sort. From the settlement agreement section 3.4d:
Except as provided in Section 3.5, AMD agrees not to participate in any Administrative Action that concerns, discusses or relates to conduct, events or
allegations that occurred prior to the Effective Date,
**** except as may be explicitly required by applicable law, such as responding to a lawful subpoena or other process administered in that Action. ****
(emphasis mine)
So much for ‘under any circumstances’.
Since Bob says it I now know it won’t happen.
Thanks Bob!
[...] de los Conspiradores está muy activo esta mañana delante del brasero comentando el artículo de Robert Cringelyen el que especula si las intenciones de Intel tras su decisión de impedir que Nvidia acceda al [...]
[...] I, Cringely » Blog Archive » Intel Will Buy nVIDIA – Cringely on technology. Related News & Resources NVIDIA, AMD Stocks Rally as Intel Larrabee Retreats [...]
[...] the past week, Intel cancelled their Larrabee GPU project. Larrabee was supposed to be a competitor to nVIDIA and ATi. ATi is a part of AMD Corporation, Intel’s biggest competitor. Intel wants [...]
Interesting theory. However I kind of think that even if they decided to merge today, they would be two chip architectures into the future by the time the deal closed, changing all the dynamics that are framing your argument. In addition I don’t think AMD or ATI is a better competitor since the merger.
[...] I, Cringely » Blog Archive » Intel Will Buy nVIDIA – Cringely on … [...]
Yes! Intc WILL buy NVDA in a couple of years and here the Great Buana reveals the reason why: AMD will have Fusion out by this time and INTC will have nothing which can compete with it. It also has no SOI-technology which will be another big advantage for AMD. And NVDA? Jen Hsung Huang just confirmed another time that NVDA has no x86 cpu or apu in the pipeline, surrendering in the chipset market, and not only the Great Buana knows that their GPU-roadmap is going to be a tragedy. NVDA is busy in renaming their 200 product line into 300 just to have something that sounds competive. Fermi? The Great Buana does not believe that there will be many monsters and other enemies going to be killed on this platform which is too late, too slow and too expensive to produce. NVDA lost contact in the GPU-market, all they can deliver is a scientific pu for a few eggheads, and something for the RISCful Tegra market which is a very crowded and highly competitive market too. Just wait and see! In a couple of years, INTC will be very happy to buy NVDA for a very cheap price!
s/buana/bwana/ (took me a while to get it. ‘Bwana’ is ‘white boss man’ (or something) in Swahili).
Mr. Cringley, this article is so bad I could have written it.
Many of the assumptions in this article — that Intel would ‘like’ to use Tegra, that Intel wants mobile GPUs, etc… are far from the truth. Intel has been on the warpath against ARM technologies. Intel dropped ARM support ages ago, and would like to keep it that way. That is why Intel signed up with Nokia.
Also, Intel is currently the graphics chip market leader with over 50% market share. Acquiring NVIDIA would put them squarely in antitrust crosshairs, with over 75% market share. It is simply a dream to believe that NVIDIA would split up their company for the priveledge of being owned by Intel.
It follows that the only logical reason Intel would want to but NVIDIA would be to shut down Tegra, not to support it. It is highly unlikely that NVIDIA management would go for that…
To summarize: the reasons you state in your blog post are exactly the opposite reasons why Intel would buy NVIDIA.
nVidia’s CEO has publicly stated on record that he does NOT want to be purchased by Intel. And he’s said this year after year. And there’s no way in hell such a purchase would clear regulatory hurdles, which is the reason why Intel had to develop Larrabee in the first place.
Might I also add that it would give Intel a combined ~80% control of the GPU market. I know the US Department of Justice has slacked sometimes, but even they’re not so stupid as to allow a merger of that magnitude.
It’s just not happening.
[...] Intel will buy Nvidia y CHW AKPC_IDS += "1008,";Popularity: [...]
Regulations? I am not sure that INTC market share in integrated matters here. Everybody knows that intels integrated graphics are not competitive in terms of gaming. And what about S3? I think that NVDA lost the opportunity to buy VIA and S3 two years ago or so. NVDA could have been there with a competivive product against Atom and Bulldozer and perhaps even the opprotunity to enter the X86 space. The Via nano has ECC and X64! But now this window is closed for NVDA and I see no chance for NVDA to escape from this mess! They have no competitive DX11 product and Tegra and Tesla do not earn enough money!
I don’t think Intel can buy Nvidia even without Larrabee for anti-turst reasons but Apple is stirring the port saying that they definately don’t want Intel IGP’s so perhaps an Apple-Nvidia behemoth could be on the cards. A great fit in fact with tegra and ion inside Apple.
How full of it are you? im just curious.
Because really… your story shows no proof of why Intel would by NVDA nor how they would get around the regulatory hurdles considering their FTC investigations.
Also, ION is not a SOC. Thats Tegra. Also, you forgot that intel isnt killing larrabee but merely not releasing that product to the public. What i foresee Intel doing is using what they learned from larrabee to make a better GPU.
[...] [...]
[...] εξαγοράσει την Nvidia.Έναυσμα για αυτή τη φήμη έδωσε το blog ‘I, Cringely‘ του Robert Cringely, άλλοτε αρθρογράφου του περιοδικού [...]
This is lame prediction. Any company can be bought if the price is high enough.
What’s the price for NVIDIA and will INTEL be ready to pay it?
Intel buying nVidia?
that’s a lame joke like what they said back then when AMD bought ATI.
even if Intel does have a chance to buy a chunk of nvidia.
I don’t think they would have a chance to beat the AMD’s Fusion chipset when they release it.
NVidia is too expensive for intel especially after they stepped on Sun’s balls.
It is true that they require a mobile chip but it is enough to lincense the brdge to nvidia they don’t need to buy it.
Nvidia will it their big ego because they need to be in x86 bussiness and intel gets what it wants.
another point is the SCC they just launched.
They prepair the terain for HPC big time by letting the universities do the research for them.
and to finish nvidia can always embrace VIA and they don’t need Intel.
In the end Jen sun wants to prove that nvidia can leave without intel and with the advent of OpenCL they can offload all that VIA cpu’s can’t on NVidia GPU-s in an very elegant manner through the driver.
And there is one more thing if they absorb nvidia it will take at least 2 years before the fruits of this embrace can be seen.
So this choice is too expensive and it takes too much time bring fruits.
by that time AMD will be sailing big time and intel will smell the smoke coming from their asses.
Compared to AMD ATI it is a completly different story.
AMD had no chipset development. NVIDIA and ATI were doing this for them.
AMD had no powerfull processors like itanium for HPC.
So to put things short there was no overlapping.
But of course this is just my opinion and I don’t know everything.
[...] Tip iCringely) Intel this week canceled Larrabee, its proposed graphics processing unit (GPU) that was intended [...]
lol sorry for the stupid mistakes:
NVidia is too expensive for intel especially after they stepped on Sun’s balls.
It is true that they require a mobile chip but it is enough to lincense the brdge to nvidia they don’t need to buy it.
Nvidia will eat their big ego because they need to be in x86 bussiness and intel gets what it wants.
another point is the SCC they just launched.
They prepair the terain for HPC big time by letting the universities do the research for them.
and to finish nvidia can always embrace VIA and they don’t need Intel.
In the end Jen sun wants to prove that nvidia can live without intel and with the advent of OpenCL they can offload all that VIA cpu’s can’t on NVidia GPU-s in an very elegant manner through the driver.
And there is one more thing if they absorb nvidia it will take at least 2 years before the fruits of this embrace can be seen.
So this choice is too expensive and it takes too much time to products.
by that time AMD will be sailing big time and intel will smell the smoke coming from their asses.
Compared to them AMD-ATI it is a completly different story.
AMD had no chipset development. NVIDIA and ATI were doing this for them.
AMD had no powerfull processors like itanium for HPC.
So to put things short there was no overlapping.
But of course this is just my opinion and I don’t know everything.
I believe that the need for Intel to buy NVidia actually comes from a cultural problem that goes back a long time in history.
There used to be this company named Evans&Sutherland (there is still such a company but it isn’t much like the E&S of old.) They produced the highest performance 3D graphics hardware in the world. SGI was started by a grad student of Dr. Evans’. Adobe was started by an engineer who worked at E&S. (I was told the first versions of postscript were developed at E&S and that E&S lasted as long as it did because of money it made from IP relationships with SGI and Adobe. True or not, I don’t know. It is what I was told.) E&S built data flow machines. 3D rendering is naturally suited for data flow machines.
As E&S started its slide into irrelevance (Ivan Sutherland had left long before and Dave Evans was sliding into Alzheimer’s disease, the management thrashed around madly trying to figure out how to get back to their former revenue numbers. (they never managed it…) One thing they tried was selling the company to Sun. The way I heard it was the last words heard from Scott McNealy were “Why would I buy this company when I can hire its engineers?”. (BTW the quotation is at best third hand so it is certain to be inaccurate and may be apocryphal.) Another thing they tried was letting a group of young engineers design a new super graphics rendering system. That system was based on a cross bar switch. The design failed because they could not solve (among other things) contention problems on the cross bar. Bus based systems and ring networks were knocked out of contention after a back of the envelope computation. Hypercubes were examined, they also failed the giggle test. (Though later I learned that Fujitsu built a hypercube based rendering system. It did not pass the giggle test
Well, E&S had a series of layoffs which naturally lead to a lot of people leaving that E&S want to keep. Most of the proponents of the cross bar system were laid off and were hired by Sun, and Intel. Over time several other key engineers were cherry picked by SGI, Sun, and Intel. For some reason, key proponents of the cross bar wound up at Intel. The rest wound up at Sun and SGI. SGI died and in the process passed their patent portfolio and many key engineers to NVidia. At least that is what I remember… A search of Linkedin.com seems to support my theory. I found job histories of enough of each group that show them moving from E&S to NVidia and Intel.
Cross bar folks at Intel, data flow people at NVidia. I don’t have any information about the internals of NVidia’s chips, but they sure act and perform like variations of the data flow machines that E&S built many years ago. (And, I just remembered that the last big E&S graphics box I read about used over a hundred NVidia chips and I know the E&S folks loved them.
Twenty some odd years later Intel announces a ring network based graphics chip to compete with data flow architecture companies in a market that has been dominated by data flow machines since the ’70s. That design, dies from the same problem that killed the cross bar at E&S way back in the late ’80s. Seriously folks, anybody who can read knows that the rendering paths in graphics APIs are called pipelines for a reason. Anyone who looked at the programming models for all the different GPU languages can see they are designed around a data flow model. What was Intel thinking? Maybe they were blinded by a cultural bias instilled by people who wanted to prove their idea *was* valid.
Ok, my bet is that it was a cultural problem introduced by hiring the laid off cross bar proponents from E&S way back when. It is possible that Intel doesn’t have a single engineer who actually understands 3D graphics. It is possible that they just didn’t bother to simulate the system using realistic data. It is more likely that Intel is the 8,000 pound gorilla that believes in its own mythology. I mean, do you remember the 432? How about the Itanium? Every so often Intel does shoot itself in the foot with a machine gun. But, maybe, just maybe, it is the effect of hiring people with a chip on their shoulders way back when.
Who cares anyway? Even if this is true — and it’s not — intel and nvidia are fighting over a stagnant market. The PC market has been so boring now for at least 10 years. There’s no reason to upgrade an older system, because the internet has made everything i/o bound by a slow network.
Serious gamers have moved on to xbox and other dedicated devices. How many people are still paying $400+ for the latest/greatest gpu? That is so 1990′s.
Apparently you should care. If not for ATI and nVidia you would not have xbox/wii/ps3 graphics. They take elements of their latest/greatest gpus to provide the graphics engines for your favorite console. If not for the high margins obtained from selling graphics cards for PCs there would not be enough money to fund the fast-paced development that trickles down into everything else.
http://www.tgdaily.com/business-and-law-features/45000-peddie-intel-will-qneverq-buy-nvidia
Not Happening. Here’s what I wrote on the City of Heroes forums when I was linked to this:
I somewhat doubt this. Not will all of the bad blood between Intel and Nvidia.
In addition, right now Nvidia pretty much offers Intel nothing of value. The Nvidia memory controllers aren’t as fast or as efficient as Intel’s. Tegra’s long-term mobile value is still iffy. Due to the recent settlement with AMD, Intel has access to all of AMD’s patents and Intellectual property, so it’s not like Intel’s hurting for graphics technology. Yeah, Larrabee didn’t make the performance cut Intel was aiming for, but since Intel is already sitting in the number volume seller of GPU’s with it’s integrated set-up, they shouldn’t be in a hurry to enter the discreet GPU market.
The only value Nvidia could bring Intel is a mass contract deal. Let’s say that Nvidia has actually captured a contract for the next Nintendo Handheld, or has captured the contract for the next Apple Ipod / Iphone system. Unlikely, but that’s the only thing that would make Nvidia an attractive acquisition target for Intel. There is also the aspect that if Nvidia had captured such a contract, they wouldn’t be in a hurry to sell themselves off to the highest bidder. They’d remain independent. So even if Nvidia did have a contract on the way, it’d still would not help an acquisition.
[...] για αυτή τη φήμη έδωσε το blog ‘I, Cringely‘ του Robert Cringely, άλλοτε αρθρογράφου του περιοδικού [...]
[...] 짜잔. 여기서 등장하는 새 떡밥 하나. … 언젠가부터 (라라비는) 엔비디아를 향한 무기가 아니라, 인텔이 [...]
[...] couple of days ago, over at his eponymous website, computer industry prognosticator Robert X. Cringely stated that nVidia will be bought by Intel, [...]
Certainly got a lot of airplay Mr Cringley. With even a modicum of ‘research’ though you would have realised Intel do not need Nvidia for graphics, they have access to all the graphics/video solutions they will need from Imagination technologies ( a smallish British IP company ) – already inside the Z-series atoms and shortly to be inside the Moorestown, then Medfield cores (as well as Canmore and Sodaville BTW).
Imagination basically own the 3D HW graphics mobile market (MBX and SGX volumes really do dwarf those of Tegra) courtesy of Apple, TI ,Renesas,NEC,Samsung etc (see http://en.wikipedia.org/wiki/Powervr for full info).
Intel own 16% of Imagination, with Apple owning another 10%, so no-one can buy it, everyone benefits.
Intel can still be seen as crushing a competitor in several markets. Chipsets, physics solutions (PhysX vs. Havok), and if long standing, long denied rumors about nVidia attempting to make a CPU are true, then add CPUs to that list.