Slashdot Mirror

← Back to Stories (view on slashdot.org)

Intel Ditches Mobile Phone Processors

Posted by ryuzaki0 on from the putting-blinders-on dept.

An anonymous reader writes "Intel is planning on selling off their XScale applications processor and 3G processor businesses for around $600 million to Marvell. From the article: 'Marvell is best known for its NIC (network interface card) chips, including wireless chipsets, and for other embedded, network infrastructure, and storage processors. The company has not previously competed in the market for mobile phone chipsets. However, it says it knows how to produce chipsets for high-volume consumer applications, which it has done for 11 years. Marvell earlier this year acquired a UT Starcom business unit in China that is working on mobile phone processors.'"

104 comments

  1. In other news by Eightyford · · Score: 5, Funny

    DC just bought AMD.

    --
    Religion for nerds. Stuff that really matters
    1. Re:In other news by Anonymous Coward · · Score: 0

      Now, if you had said DC Comics just bought AMD - that might have been funny...

    2. Re:In other news by Fordiman · · Score: 2, Funny

      Naw. Instead, he dropped a joke grenade. Wait ten seconds, THEN laugh.

      --
      110100 1101000 1101000 1100110 0 1101111 1101000 1100011 1
  2. Headline is stupid by drinkypoo · · Score: 4, Informative

    ...and cribbed directly from the article - where it was also stupid.

    XScale is not, repeat not a "mobile phone processor" although I'm sure it's used there. In fact they specifically sold the PXA line, which includes the processor in my iPAQ.

    It never ceases to annoy me when someone is so lazy that they can't even write their own headline - especially when it's wrong. If you're going to plagiarize, why not copy something that's actually correct?

    --
    "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    1. Re:Headline is stupid by bsdluvr · · Score: 1, Troll

      In Soviet Russia, plagiarism edits slashdot.

    2. Re:Headline is stupid by nerdyH · · Score: 5, Insightful

      Have another drink, drinkypoo, and make more blather about nothing. H/L is accurate. Intel also sold baseband phone processors to Marvell. Though PXAs are used in PDAs, mobile phones are probably 95 percent or more of their volume, I'd guess. Intel did not sell the whole XScale line... just the xscale's that go into phones.

    3. Re:Headline is stupid by hawkbug · · Score: 1, Redundant

      Well, the Xscale is used in the Blackberry 8700G - which is a mobile phone / PDA.

    4. Re:Headline is stupid by Dun+Malg · · Score: 1

      Well, the Xscale is used in the Blackberry 8700G - which is a mobile phone / PDA.

      I contend that if it looks like you're holding a pocket calculator against your head, it's not really a mobile phone.

      --
      If a job's not worth doing, it's not worth doing right.
    5. Re:Headline is stupid by Shadyman · · Score: 1

      Because who's ever heard of a PDA doubling as a mobile phone? That's crazy-talk!

    6. Re:Headline is stupid by hawkbug · · Score: 1

      Have you actually seen the 8700g? It's not much wider than a standard phone - and it's the same height and not thicker.

    7. Re:Headline is stupid by HeX314 · · Score: 1

      But it's the kind of crazy-talk you'll have 800 minutes of every month.

    8. Re:Headline is stupid by drinkypoo · · Score: 1

      I realize that English is a difficult and convoluted language, but if you say "mobile phone processor" that means a processor used only in mobile phones. It really is just that simple in this case. Sometimes it's hard, sometimes people just want to be obstinate and defend misuse of language. What really amazes me is that programmers will so frequently defend such behavior, when they should really be the ones who are most aware of the need for precision. Change a word, move a word or two around, and the meaning changes completely.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
  3. Doesn't seem like a big deal. by celardore · · Score: 3, Insightful

    Many mobile phone producers have their own completely adequate chipset solutions. I'm not sure how many cellphone producers rebrand chips they use though. I am sure that if a phone provider needed Intel hardware for whatever reason, they could simply buy and or rebrand the chips or rights from Intel if the need arises.

    1. Re:Doesn't seem like a big deal. by drinkypoo · · Score: 1
      yeah, they could, unless it was a PXAxxx type processor more than two years from now, in which case they'll have to buy it from Marvell :P

      Some people don't mess with intel stuff though... AFAIK all the motorola phones (for example) are based on motorola chips.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    2. Re:Doesn't seem like a big deal. by treeves · · Score: 2, Informative

      RIM is also using the Intel Xscale chips in the Blackberry.
      From TFA on red-electronics.com, Intel will keep making the chips for Marvell until Marvell finds another manufacturing solution - probably TSMC or the like, my guess is.

      --
      ...the future crusty old bastards are already drinking the Kool-Aid.
    3. Re:Doesn't seem like a big deal. by corychristison · · Score: 1

      Perhaps it is because, as he_humeister mentioned, that they are losing money in the area.

      I also think it may be because of the nice tight relationship with Intel now-a-days... Just a thought. :-)

  4. What Happened to Diversification? by ewhac · · Score: 4, Insightful
    Wow, that Pentium basket must be awfully durable for Intel to be putting all their eggs in it. Or maybe Intel prefers not to be in a market in which there are about a dozen players (namely, providers of ARM-based system-on-chip products).

    Schwab

    --
    Editor, A1-AAA AmeriCaptions
    1. Re:What Happened to Diversification? by Anonymous Coward · · Score: 0

      I dunno I think they were more afraid of being investigated for duopoly with Microsoft back then that they needed something to fall back on incase they were forced to do something at behest of DoJ about the processor line. They probably figure that since they now have decent competition for desktop, server, and notebook PC chips, and since they need to shore up their business, it's convenient for them to sell off this line of chips.

    2. Re:What Happened to Diversification? by the_humeister · · Score: 1

      Or maybe it's because they've been losing lots of money on the unit. If they'd actually made money on it, you'd think they'd keep it. Although then you'd wonder why Itanium is still around...

    3. Re:What Happened to Diversification? by Fordiman · · Score: 1

      Query:
      Do you know of a powerful (> 1GHz) and inexpensive (< $200) ARM mobo/chipset solution?

      --
      110100 1101000 1101000 1100110 0 1101111 1101000 1100011 1
    4. Re:What Happened to Diversification? by Anonymous Coward · · Score: 0

      Some (or all?) of Intel's Xscale procs are ARM-based.

    5. Re:What Happened to Diversification? by automatix · · Score: 2, Interesting

      Spending the last couple of years working for a consumer of XScale processors, I can tell you Intel's XScales are severely lagging behind Freescale, Samsung, and others. The Intel chips are slower, have less features, and are more expensive than the Freescale and Samsung offerings. Large consumer electronics customers are dropping them, especially as Samsung will offer better deals on processor+flash+ram offerings than Intel can. Because they're all based off the same ARM cores the application-level software shouldn't need to change much, and it is a viable choice when creating new hardware platforms. Intel does have nice marketing agreements where they pour 10s of millions into companies each year to put 'Powered by XScale' on the packaging, advertising, and software.

    6. Re:What Happened to Diversification? by LordMyren · · Score: 2, Informative

      Intel really just pulled a Pentium IV with XScale. It was basically a StrongARM core jacked up in mhz. The new ones run like 700 mhz, but they're the same architecture as the 100 mhz StrongARM they ran a decade ago. Very poor integration, power consumption not that great, just not a good chip, except if you look at the mhz. Like you say, they kind of got their pants beat.

    7. Re:What Happened to Diversification? by feijai · · Score: 1
      The new ones run like 700 mhz, but they're the same architecture as the 100 mhz StrongARM they ran a decade ago. Very poor integration, power consumption not that great, just not a good chip, except if you look at the mhz. Like you say, they kind of got their pants beat.

      I guess that explains why everyone prefers them in the PDA and mid-size embedded market then. And I mean everyone from Gumstix to Palm TX to Dell Axiom.

      I have no idea what you're talking about power consumption wise: can you name another chip which has 0.001 W/MHz and runs at at least 100MHz? I can't. And the current XScale package used on the Gumstix seems to have pretty good single-chip all-in-one integration. No USB2 yet, I guess.

      Thing is, though everyone in that market uses XScale, it's a bad market. My guess is that Intel has realized that the StrongARM architecture was designed for PDAs (specifically, the Newton), and PDAs are a rapidly shrinking market. The chip consumes too much power for cell phones and they don't need its speed, nor do routers etc. And it's not powerful enough for PCs or servers. It's probably just a chip that's king of the hill in a shrinking market.

    8. Re:What Happened to Diversification? by Ed+Avis · · Score: 1

      No they're just waiting until they launch the low-power, embedded Itanic.

      --
      -- Ed Avis ed@membled.com
    9. Re:What Happened to Diversification? by mrchaotica · · Score: 1

      Are you asking because you specifically want that instruction set, or because you want a fast, cheap, and low-power-consumption chip? If it's the latter, the AMD Geode might interest you (it's an x86, though).

      --

      "[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz

    10. Re:What Happened to Diversification? by MemoryDragon · · Score: 1

      I cannot comment on the technical background of the XScale except that I constantly had the feeling the design was rather old. You constantly could read about new ARM generations, new extension modules like a hardware java acceleration etc... yet XScale was XScale no new features only a bunch of more MHz while others moved towards newer ARM cores. The reason why you can find mostly XScales in pdas probably is, the intel brand recognition on those devices, speaking of windows + intel, which basically pushed many wince makers towards xscale although neither wince nor the intel xscale were any compatible to their pc counterparts. And of course, back in the early days of WinCE the XScale was one of the fastest ARM based processors around, this advantage basically was kept up about one year and was a result of the DEC takeover, Intel inherited basically what the Alpha people at dec did with the ARM core and screwed it up majorly, by not pushing the technology really further. I constantly had the feeling the XScale was the unwanted stepchild of intel, there, because it sold in pdas, but not really wanted, because it was no x86 and shock not even an intel design and even worse, the core was licensed from outside (no matter that ARM are still currently the best processor designers in the area of low power embedded procs)

    11. Re:What Happened to Diversification? by Fordiman · · Score: 1

      Feh. I know of plenty of embedded cisc chips. I'm specifically after a relatively powerful ARM, but the best I've found is a gumstix.

      --
      110100 1101000 1101000 1100110 0 1101111 1101000 1100011 1
  5. Marvell by Ramble · · Score: 0

    So when is there going to be a teen romance story involving superpowers between Paul Otellini and Gordon Moore?

    --
    "Oh boy"
    1. Re:Marvell by Anonymous Coward · · Score: 0

      Where is Trollaxor when you need him?

  6. Marvell? by theheff · · Score: 2, Funny

    I thought cell phones powered by standard CPU chips was only something you see in comic books...

  7. lost billions of dollars by vivek7006 · · Score: 5, Informative

    Intel has lost billions of dollars since late 90s on this. EE-times gives some more details http://www.eetimes.com/news/semi/showArticle.jhtml ?articleID=189602065

    During the course of the past decade Intel invested between $3 billion and $5 billion in the assets it sold to Marvell, says Will Strauss, an analyst for Forward Concepts. Intel spent nearly $2 billion on a single acquisition to bolster those communications chip efforts. It was a major rat hole of unparalleled magnitude.

    1. Re:lost billions of dollars by Short+Circuit · · Score: 5, Interesting

      It was a major rat hole of unparalleled magnitude.

      How much did they spend on Itanium, again?

      --
      tasks(723) drafts(105) languages(484) examples(29106)
    2. Re:lost billions of dollars by DAldredge · · Score: 5, Informative

      Between 7 - 12 Billion USD.

    3. Re:lost billions of dollars by Anonymous Coward · · Score: 2, Insightful

      Everyone makes fun of the Itanium, but it isn't like all of the money that was thrown into it was wasted. First and foremost, it did its job: it got rid of many competing companies. Sure, that was before it was actually released - but damn, it looked good on paper. Various companies decided that they could never compete with Intel, and they cancelled their equivalent product lines.

      Second, it isn't like all of the money that was thrown into the chip's design is wasted. Itanium was effectively a testbed for many new technologies that are now found in standard Intel chips, be it the P4, their mobile lines, or the new Core 2 Duo.

      Certainly, the Itanium is not a standard "we made money" success - but it isn't as bad as its made out to be (at least, in terms of Intel's use of it. For the actual consumer, it wasn't a good thing)

    4. Re:lost billions of dollars by janet-on · · Score: 0

      Motorola "spun off" (ie: ditched) their chip-making business. Inmos - owned by a music chain, Thorn EMI - was sold to ST and their technology was dumped. IIT, a co-processor manufacturer in the days of the 8086 to 80286 died a death. Cyrix was bought, as mentioned.

      This is a field where you must not only have a good product, you must also have a solid market AND a solid marketing team, AND you must avoid bad PR like the plague, AND any major players (like Intel) must not deliberately sabotage efforts to compete, AND your plant can't be struck by major earthquakes.

      (Why are all the major chip makers in Taiwan, Japan and America ALL concentrated in areas with high tectonic activity? Is there something in the fault line they use in the production line?)

      The bottom line is simple. A chip fabrication plant can cost tens of millions to hundreds of millions of dollars, skilled chip designers can command hefty salaries, many of the key markets are 0wn3d by monopolies of questionable legality who flirt with unethical practices to keep their position, and software developers reinforce this by targetting established, high-volume platforms and that means no new products get support.

      Of course, Transmeta didn't help its case. Its Linux distro was late, the first batch of chips was buggy, they didn't sell to anyone outside of the "big players" (and "big players" only really buy from other "big players", because volume bought and sold = profit), and they only produced an 80x86 layer for the Crusoe, rather than using the capabilities to cross market boundaries and therefore create volume by getting into many niche markets.

      Also, their design was poor. Intel beat them on power consumption in a very short space of time, and this is Intel we are talking about. At the same time, people knew there were problems with 80x86 scalability (hence the work on SMP and hyperthreading), but Transmeta didn't look far enough ahead to build a multicore product, when they were already building a design from scratch and had ample opportunity to make such changes.

      (In comparison, AMD and Intel have to engineer such features into an existing design, which is always much harder and likely to be much slower than working from first principles. AMD's and Intel's route also offers much better odds of bugs being found in the design, at a later date, as their architecture was never intended to be multicore.)

      So, I don't hold Transmeta blameless in this. They may have been pushed over the edge, but they still chose to walk along the cliff in the first place, knowing it to be a dangerous spot, and knowing that the view wasn't even that good there, to make it worth the risk.

      One of these days, I hope to see a company start up that takes the time to be truly innovative (and not just fake it), takes the time to get things right, and makes a product so damn unbeatable it wipes the floor with everything else.

      It does happen. True, AMD is no start-up, but they were hardly giants in the 80x86 world. With the Opteron and their 64/32-bit crossover architecture, they've demolished Intel's Itanium and even convinced Microsoft to switch to them for 64-bit stuff. Given the longevity of the Wintel duopoly, that took a good plan and a good effort.

      Any start-up could do just as well, or better, because it wouldn't have the legacy hardware to build around. They could do a clean design that merely supported legacy code. Transmeta started down that road, but for some reason chose only to camp a little way down it and go no further.

      The "ideal" processor would work just as well as a CPU, GPU, network processor or processor for a disk array, as then a manufacturer can go to a single vendor, buy in even bigger bulk, and save money on all aspects. Your computer would become a Beowulf cluster, in effect, with specialization in software. It would be cheaper to build, and would mean that the same system would work for a desktop or a server. You'd simply load in different specializing software, to shift the processing

    5. Re:lost billions of dollars by TheKingAdrock · · Score: 1

      Itanium was not a testbed, and besides perhaps some low-level circuit design aspects, there is nothing from Itanium in the other chips - nothing architectural, that is. Apparently you're talking out your...

    6. Re:lost billions of dollars by SchwarzeReiter · · Score: 1

      Its good to see that old Intel is able to lose money even on a market which is actually growing.

    7. Re:lost billions of dollars by Anonymous Coward · · Score: 0

      Yes, this is true. The other reply to the parent post is nothing but trolling. There are a number of situations where Intel can take what it has learned in Itanium development, and use that in developing x86 processors (and yes, before someone states the obvious, this holds despite the fact that the architectures are radically different). This will be particularly relevant when Itanium gets yet another major refresh a couple of years from now. Everyone makes fun of the Itanium simply because it didn't go fit along with Intel's ridiculous expectations. From a technical standpoint, the latest 1.7 billion transistor Montecito core is nothign to be made fun of. Then again, wow-gee-whiz figures are lost on most slashdot readers, since most of them have never tried developing a piece of technology so complex... Wow, this was a long post.

  8. Marvel? Perfect! by mottie · · Score: 1

    Now to figure out how to pick the cellphone that has mutated and has special powers..

    1. Re:Marvel? Perfect! by Anonymous Coward · · Score: 1, Insightful

      And to sign our lives and first born children away just to get the programming specs!

    2. Re:Marvel? Perfect! by Anne+Thwacks · · Score: 1
      Like making video calls, playing MP3s, playing games?

      Hell, My landline phone cant even remember its phonebook through a power cycle, and anyway the phone book is so un userfriendly, its never had more than one number in it.

      I think both my mobiles already ahd supoer powers. My Nokia now even has "Frogger" - I am just waiting for "Lounge Suit Larry"

      --
      Sent from my ASR33 using ASCII
  9. Cute joke, but... by Valdrax · · Score: 2, Insightful

    I'd hardly call a super-low power consumption embedded processor without a floating point unit a "standard CPU chip."

    --
    If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
    1. Re:Cute joke, but... by LordMyren · · Score: 1

      The new phillips 90nm lpc3180 has not only floating point, but vectorized floating point, and some of the lowest power consumption available. Mono runs great on it. That seems pretty standard to me. ARM really is the new first class citizen of the CPU world, long after x86 pushed everyone else out.

  10. An event all too familiar... by mantar · · Score: 4, Interesting

    Interesting... I also heard that Intel is looking to off-load their telecom subsidiary, Dialogic. I wonder what's going on with these guys?

    --
    # man tar
  11. Quick, go update the museum! by Anonymous Coward · · Score: 0

    I visited the Intel museum for the first time on Saturday. They have a prominent exhibit showing all the mobile devices Intel makes chips for. I guess they're going to need to do some quick exhibit shifting.

  12. The IP Value is in the HDL Code, Not In The Chip by loose+electron · · Score: 1
    Any significant cell phone vendor has its own digital ASIC team doing a fully dedicated logic ASIC. Impossible to be feature (or cost!) competitive by doing this any other way.


    The alternative model is Qualcomm which develops a full chip set (everything, RF front end to the Sigma-Delta ADC/DAC that drive the speaker and hear the microphone.)which then gets bundled into assorted CDMA phone sets.


    These chips don't require cutting edge speed, rather they are totally cost vs. feature driven. (How many toys and games can you pump into your digital ASIC and firmware for the best cost/feature tradeoff?)


    Be it in-house digital ASIC team, or the "end to end bundle" of Qualcomm, they still will have to have the digital ASIC ready and able for easy customization. (Samsung phone playing Pong and Toshiba phone playing anime' graphics, so to speak.)Consequently, the value is in the Verilog/VHDL code, and the capability to make additions, so that there is one and only one logic ASIC. That's a "must have" due to cost-power-size issues, all critical in cell phones.


    If Intel thinks they can sell a chip for this that fits the "Intel motherboard and chip set model" it won't happen. Considering they are dumping this group that is bleeding red ink, that is not too surprising.

    --
    www.effectiveelectrons.com "chips that work" Analog, RF, Mixed Signal
  13. Apple to Intel theories by Anonymous Coward · · Score: 0

    Wasn't one of the theories about Apple switching to Intel that they'd have a vendor able to provide both their PC and Ipod CPUs, and in doing so qualify for a bulk discount? That while they're shipping a lot of CPUs if they were able to twice as much any vendor would have to give them more of a discount? This was the theory posted on Arstechnica by Hannibal and linked here.

    Guess it shoots down that theory.

    1. Re:Apple to Intel theories by l33td00d42 · · Score: 1

      Wasn't one of the theories about Apple switching to Intel that they'd have a vendor able to provide both their PC and Ipod CPUs? ... Guess it shoots down that theory.

      ... unless the next ipod is going to draw 200 watts and be powered by an RTG. ;)

  14. Wallstreet fashion driven by EmbeddedJanitor · · Score: 3, Insightful
    Most/many decisions are not driven by a sane business plan, but by the latest Wall St fashion. If you're not doing "it" (whatever "it" is), then you get punished on Wall St, particularly if your stock is looking a bit stagnant/down. So industries follow these trends: diversification, refocussing on core business (divestment), off shoring, sigma 6, whatever.

    This quarter's fashion seems to be divestment.

    Anyway, Intel were not making much money (??were making a loss??) on their PXA line. The PXA plays in a highly competitive market with a lot of players (TI, Samsung,...) and very little brand loyalty (No Intel Inside message). Intel has never held up well to that sort of competition and have got out of many businesses when things got hot (RAM, 8051, USB chipsets,...).

    --
    Engineering is the art of compromise.
    1. Re:Wallstreet fashion driven by billcopc · · Score: 1

      What is the opposite of brand loyalty ? I cringe whenever I see "Intel Inside" because to me that's a warning of all the stupid glitches and thermal issues I'm going to be dealing with. Chipsets spontaneously combusting and setting the internal cabling on fire. CPUs throttling down because the heatsink can't cope with the P4's horribly inefficient high clockspeeds. Video getting garbled because aforementioned heatsink is blowing 150' hot air from the CPU to the board/graphics chip, pressure-cooking the bus.

      Intel Inside... run for your lives!

      --
      -Billco, Fnarg.com
  15. OK forget what I said... by Bruce+Perens · · Score: 2, Insightful
    In a story here a month or so ago about Intel abandoning their embedded business (which I can't find because the search function isn't working), I wrote that they were doing no such thing as long as they held on to ARM (XScale).

    Now, we see they're not.

    Hm. Lots of eggs going into only one basket. Is this because they took a financial hit on Itanium?

    Bruce

    --
    Bruce Perens.
    1. Re:OK forget what I said... by Scott7477 · · Score: 1

      I have been wondering who at Intel would be getting fired as a result of this SNAFU, and stumbled on this commentary http://techsearch.cmp.com/blog/archives/2006/06/do es_intel_xsca.html?loc=hardware/
      that suggests that Otellini is actually buying himself some time with this sale.

      --
      "Lack of technical competence coupled with the arrogance of power, as usual, leads to no good end."
    2. Re:OK forget what I said... by chrish · · Score: 1

      So much for my theory about Apple going with an Intel exclusive to get a sweet discount on embedded processors for the iPod line...

      --
      - chrish
    3. Re:OK forget what I said... by bill_mcgonigle · · Score: 1

      Yeah, I expected an OSX-lite iPod on XScale as well, but I think pricing pressure precludes feature-heavy iPods. Maybe since Apple's going with the competitor to PortalPlayer for their new units, that marked the end of the Intel chance to supply the iPod parts.

      --
      My God, it's Full of Source!
      OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
  16. Err, what? by mcbridematt · · Score: 1

    You would think with just about every PDA on the market being XScale powerered, that they would be making tons of cash, at least on that...

  17. Surprising... by Anonymous Coward · · Score: 0

    Looking at the specs of different PDAs - iPAQs, Palms, Treos, Blackberries and even the new Motrola Q use Intel XScale processors. And yet the division lost the company billions of dollars.

  18. Itanium to follow suit? by edxwelch · · Score: 1

    This is the result of the pressure that AMD put on Intel. They can no longer afford to spend money on loss-making operations. It would make sense if they got rid of the Itanium as well. Surely it must be redundant now that they have Woodcrest.

  19. Because.... by GillBates0 · · Score: 0

    ...only hindsight is 20/20. Foresight is 0/20, if that.

    --
    An Indian-American Hindu committed to non-violent thought/speech/action alarmed by the global explosion of radical Islam
    1. Re:Because.... by Deltaspectre · · Score: 1

      I'd just like to say that.... 0/20 is infinitely good vision :)

      --
      My UID is prime... is yours?
  20. Will it suck to develop for Xscale now? by Anonymous Coward · · Score: 1, Insightful

    Marvell just isn't marvelous when it comes to providing documentation. Just try to get any information on any of their chips in the cheap consumer routers.

  21. Even more interesting... by Anonymous Coward · · Score: 0

    Funny how "/." had no mention of Intel's Woodcrest launch on Monday.

    I guess stories about non-AMD processors aren't news for geeks.

  22. this probably has to do with DaVinci by Locutus · · Score: 4, Insightful

    DaVinci is Texas Instruments single chip solution for mobile phones and multimedia rich embedded devices. They mixed a TI DSP chip in with the ARM core( anyone remember OMAP ) for a high performance single chip solution. Prior to this, smartphones used one processor for the radio and one processor for the GUI/applications. The holy grail here is one processor for everything significantly reduces cost. Intel DSPs are not near as popular as TI's and so it's a no-brainer to use TI's stuff in this case.

    http://hardware.newsforge.com/article.pl?sid=06/01 /05/163242&from=rss
    and
    http://www.ti.com/corp/docs/landing/davinci/firstp roducts.html?DCMP=DSP_DaVinciCatalog&HQS=Other+PR+ thedavincieffectpr

    LoB

    --
    "Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
    1. Re:this probably has to do with DaVinci by HoserHead · · Score: 2, Interesting
      Further, I've heard a lot of people griping about (for example) the OMAP processor in the Nokia 770, which runs at 220 MHz -- saying "My xscale runs at 400 MHz, this is seriously underpowered." Sadly, this is far from the case.

      The fact is that Intel royally screwed up the xscale processor - in a past life, I worked at an embedded Linux company, and once we'd switched from a 200 MHz OMAP chip to probably a 300 MHz XScale, our performance went way down. I/O, in particular, was atrocious on the XScale.

      Add to that the fact that TI is legendary in the industry for low power consumption, and you end up with the net result that TI tends to win embedded sockets more often than not.

    2. Re:this probably has to do with DaVinci by Locutus · · Score: 2, Interesting

      I don't yet know about the Nokia 770 but I think you might have run into the PXA255 fiasco. A PXA250 running at 206MHz ran circles around the PXA255 running at 400MHz. Intel screwed up that chip so bad that cache had to be turned off in many cases and I think there was one other bug in it which also greatly reduced its speed.

      Simply amazing how Intel has blown not only the desktop CPU market but also the handheld/etc market.

      LoB

      --
      "Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
    3. Re:this probably has to do with DaVinci by Anonymous Coward · · Score: 1, Informative

      I think you means the SA1110 at 206 MHz was faster than a PXA250 at 400. The PXA255 was created to fix all the problems in the PXA250.

    4. Re:this probably has to do with DaVinci by Locutus · · Score: 1

      That's what I ment alright. Thanks.

      LoB

      --
      "Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
    5. Re:this probably has to do with DaVinci by Anonymous Coward · · Score: 0

      Lots of people complaining about the PXA250 ran their old binaries compiled for StrongArm. Since the PXA had a re-designed pipeline it's no wonder performance was poor. After recompiling you usually got a performance increase beyond the MHz difference. There was no flaw here, just lack of information.

    6. Re:this probably has to do with DaVinci by Erich · · Score: 1

      OMAP (even OMAP3) is still an applications processor, it does not do baseband processing, you need a separate chip for UMTS/HSDPA/HSUPA. The DSP and coprocessors are for imaging, audio, and video. The ARM sucks as a DSP; you're not going to be able to do H.264 VGA encode/decode in software on an ARM.

      --

      -- Erich

      Slashdot reader since 1997

  23. they killed Pentium... by YesIAmAScript · · Score: 1

    I hear you about having just x86. Seems risky. But you realize they have abandoned Pentium, right? There's only one Pentium left to be released. Everything else is Core .

    --
    http://lkml.org/lkml/2005/8/20/95
    1. Re:they killed Pentium... by Cadallin · · Score: 1

      What do you think core is? It's still a freaking Pentium Pro. It's a Pentium Pro that's had a lot shit tacked on, and now (finally) gotten a slimdown redesign, but it's still a Pentium Pro deep down.

    2. Re:they killed Pentium... by Anonymous Coward · · Score: 1, Informative

      Sure 'nuff... it's had a strange history. Core is a Pentium M with SSE3 tacked on, (x2 dfor the dual cores..),. Pentium M is based on a Pentium III M... pretty much a die shrink, P4-style bus to make the chipsets easier, and probably some other stuff. P3M was a workover of the Pentium 3 to use less power (not even done by Intel -- they hired a outside co. to do the work). At the same clock speed, the Pentium 3M would kill the Pentium 3 performance-wise, and use like 1/4 as much power. Pentium 3 already outran a Pentium 4 at the same clock.. pentium 3 was based on pentium 2, and I think probably was just a die shrink and extra instructions too... Some P2 motherboards would in fact accept the slot P3 processors and operate, as long as board supported 100mhz FSB and the BIOS would deal with it. Finally.. Pentium 2 was a Pentium Pro that sped up support for 16-bit code, plus a die shrink.. Yee-haw. The Pentium Pro ran 32-bit code fast, but Intel figured people wouldn't still be running DOS by the time it came out, so 16-bit code ran nice and slow. Well, Windows 95 & 98 are just bloated DOS shells, plus have tons of non-32-bit code from Win3.1 and previous, so they ran totally crapulently on the PPro. I think NT was supposed to run pretty good, and I've seen NetBSD and Linux haul ass on a 200mhz one. A quad PPro is definitely fast 8-).

    3. Re:they killed Pentium... by mrchaotica · · Score: 1
      Finally.. Pentium 2 was a Pentium Pro that sped up support for 16-bit code, plus a die shrink..

      ...plus MMX.

      --

      "[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz

  24. is the license transferable ? by johnjones · · Score: 1

    intel has a architecture license for ARM processors it had to buy one even after it took over the StrongARM from digital

    now what happens ?

    realistically I see more legs on a chip marketed from a company other than intel BUT

    INTEL IS going to be pushed out of alot of markets simply because it does not have a solution for them now that Xscale is not in the stable

    this is bad for intel but I suspect it makes a easy argument at exec level (well as these easy as these things can be ) because they want to be seen to do something

    forget about the emerging PDA/Smartphone market (and just when it was getting intresting with Microsofts intergarted message product and MS VOIP products )

    silly

    regards

    John Jones

    1. Re:is the license transferable ? by alienw · · Score: 1

      The PDA/Smartphone market is not "emerging", it's quite dead.

    2. Re:is the license transferable ? by MemoryDragon · · Score: 1

      It is not emerging, but being very strong, it is just that pdas are dying but you can find an arm processor in millions of phones worldwide, but not from Intel, mostly... Same goes for the automotive embedded market where power processors rule the world there, but nothing from Intel. Pretty much the entire embedded market is intel less.

  25. "Intel Architecture" by l33td00d42 · · Score: 1

    will it finally be correct to call x86[_64] the "Intel architecture" (as a bunch of knumb-nuts called it back at the Apple switch) or is IA64 still hangin on for dear life?

  26. marvell now limitless? by inexia · · Score: 1

    this potential deal could really cause a quake in the world of wireless (please throw red flags if i am wrong) hynix, infineon, and the like could really jump on board something this sweet and take advantage as Marvell would now have somewhat a threshold over the mobile environment (again, throw red flags if i am wrong). Micron was a suitor for hynix but it all fell through quite badly and infineon makes a pretty stout product. Samsung seems to integrate them well and the products come out top notch. Marvell could be onto something that may produce great products with the right integral partners. *waiting for verbal abuse*

  27. Interesting.... by niXcamiC · · Score: 1

    Does this mean that Core beat ARM, cause I always though it was the other way around.

    --
    Chances are any disscution on Slashdot will degrade into a flamewar about ID/Christianity within 14 posts.
    1. Re:Interesting.... by dingleberrie · · Score: 1

      I never realized that connection before. For those not in the know about Total Annihilation

  28. Strange but not Incomprehensible by LordMyren · · Score: 2, Insightful

    This is pretty weird news, pretty unexpected. Intel's been trying to make inroads on embedded for years, they know there's huge volume there. StrongARM and XScale were kind of their front line warriors in that battle. Presumably, they're going to be relying on convincing people to use low voltage Core's in the future. Continuing an ARM based line would only draw attention away from their amazing x86 market. It still seems flaky though, given that x86 hasnt been used as a SoC in a long time; 80186 or so. Cell phone with a north bridge, anyone?

    On the other hand, while StrongARM was a reasonable contender in the ARM market, the initial XScale models provided virtually no real enhancement over StrongARM, and often increased power consumption in the process. This was a long time ago, but I remember some rather tempermental items on the Errata sheets. Intel simply wasnt cracking heads like the silicon giant it wanted to be. It just wasnt an impressive processor in any respect. Its probably three or four years old now, and Intel's decided the experiment has come time to wind down.

    All this as newer faster better ARM cores keep showing up.

    I really want to see what Intel's next move is. I am certain they're not going to drop the embedded sector, I know they realize how big it is, how massively its growing. What they're next heading is after this move, that should prove quite interesting.

    -LM

    1. Re:Strange but not Incomprehensible by johu · · Score: 1

      Actually there's embedded x86 processors on market. Not from Intel but RDC. There's some conflicting information on web, but RDC R3211 is actually x86 and for example Linksys WRT54GR uses it and runs x86 Linux.

      There's also NS/AMD Geode. Of course not everyone agrees which use should be categorized as embedded. Probably neither Geode nor RDC is used on cell phones.

  29. Because AMD will do so by pampi · · Score: 1

    They did that because AMD will do so

  30. the future of microprocessors????? by SylvesterTheCat · · Score: 1

    1. Motorola sells / spins off its microprocessor division (ColdFire, et. al.) into Freescale.
    2. Intel sells its XScale microprocessor.

    Collectively, this makes a significant portion of the >= 16 bit microprocessor market. (Sorry... I'm being conservative here. I suspect that both are 32-bit, but since I've been drinking a bit this evening, I'll error on the side of caution... )

    What does this say for the state of (and the future of) the embedded microprocessor world?
    Are they saying that these markets are not profitable? That can't be.... can it?

    What is the future for Atmel AVR family? ... or the 8051-family?

    Hmmm... Must drink more and think about it..head hits keyboard..........

  31. 500ghz by ScaryFroMan · · Score: 1

    Must have been because they were intimidated by that chip that ran 250 times faster than a mobile phone processer that was in the news a couple of days ago...

    --
    In Soviet Russia, backwards is everything.
  32. Too bad by scatterbrained · · Score: 2, Insightful

    Xscale was one of the better product lines from a small hardware
    developer perspective - good docs, good cast of supporting tools,
    resonably inexpensive parts that could do a lot. Now it's going to
    Marvell, whose tight assedness about documentation and NDAs makes
    even Broadcom look like a bunch of free-love hippies. sigh...

    --
    -- All that's left of me, is slight insanity, whats on the right, I don't know. -- Bob Mould
  33. Welcome the Newest AMD Fanboy by Anonymous Coward · · Score: 0

    I work for the division that Intel just sold. Now I say screw em!

  34. if you're gonna say that... by YesIAmAScript · · Score: 1

    Then I can just as easily say it's all the 80386. Because these new fancy chips may have new names like core, and they may have been derived from Pentium Pros, but Pentium Pro was derived from 386, and so we all know they're all just 80386s deep down.

    --
    http://lkml.org/lkml/2005/8/20/95
    1. Re:if you're gonna say that... by mrchaotica · · Score: 1

      Bah, that's not far enough either, because the 386 was just a 286 with a memory management unit tacked on...

      --

      "[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz

    2. Re:if you're gonna say that... by WilliamSChips · · Score: 1

      Whatever happened to the 186? Why do you hear of 286, 386, 486, and 586, but never 186?

      --
      Please, for the good of Humanity, vote Obama.
    3. Re:if you're gonna say that... by tzanger · · Score: 1

      Whatever happened to the 186? Why do you hear of 286, 386, 486, and 586, but never 186?

      The 80186 was an 8088/86 hybrid with on-chip peripherals. It was intended for embedded applications. The old unisys ICONs used them.

      (that was from memory, but Wikipedia backs me up on this. God I must be old to not only remember this stuff but also the NEC v20 and MOS 65xx histories...

    4. Re:if you're gonna say that... by Cadallin · · Score: 1

      Wrong, the Pentium Pro was the first x86 chip not derived from the 80386. It had a front end designed for decoding the same Instruction set as the 80386, but at the architectural level, the two aren't even remotely related. The Pro was the first Intel chip to decode x86 instructions in micro ops (a special instruction set never exposed externally) it is these the Pro executes, and the architecture is other very different.

  35. Marvell - too bad by Alioth · · Score: 1

    Too bad it's going to Marvell - these guys are apparently pretty hostile towards open source software, not releasing specs.

    --
    Oolite: Elite-like game. For Mac, Linux and Windows
  36. Question by MemoryDragon · · Score: 1

    The XScale design was inherited by DEC, I constantly got the feeling once Intel took over they simply had it as a sideprocessor, never being happy with having a licensed core instead of their own. I did not look into the XScale development over the years too much, but did Intel ever integrate newer arm cores, or modules like the java vm extensions. I just wonder because most ARM processor manufacturers ramped up their procs with newer designs, why I had the feeling from an outside perspective that intel wanted to have this processor line die, but was surprised with cold feets once they became sort of defacto standards in all windows ce based pdas.

  37. Everybody has a PDA by bill_mcgonigle · · Score: 1

    My guess is that Intel has realized that the StrongARM architecture was designed for PDAs (specifically, the Newton), and PDAs are a rapidly shrinking market

    Huh? Everybody I know has a PDA they carry with them all the time. They also have a cellular radio built in.

    --
    My God, it's Full of Source!
    OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
  38. iPod Impact by west · · Score: 1

    There was a lot of speculation that part of the reason for Apple's switch to Intel was that Intel might offer better pricing, development, etc. on XScale, which would then find a home in new iPods. If this was indeed the case, I bet there's one steamed Steve in Cupertino.

  39. XScale not just for phones by MMHere · · Score: 1

    My Sharp Zaurus SL-C3100 (a PDA) has an XScale processor, and runs ucLinux under the hood.

    Does anyone see this change impacting Sharp's PDA products?

  40. hmm. by YesIAmAScript · · Score: 1

    I think the 8086 was the first x86 chip not derived from the 80386. It also was microcoded, like every other processor of its day. Microcode is an internal format not expose externally. 486 was the first processor in the family to break from microcoding, they "compiled the microcode" for common instructions, giving them much faster execution by putting them into fully decoded logic instead of micro-ops.

    Pentium also broke down instructions into ops, stuffing them into separate pipes. And since it had parallel pipes, it had to have a special retirement/writeback unit that was greatly different than any before (although less complex than Pentium Pro). 386 and earlier didn't even have a pipeline!

    --
    http://lkml.org/lkml/2005/8/20/95
  41. 186 was a real dud.. by YesIAmAScript · · Score: 1

    188 was a bit more successful, due to allowing lower cost designs (at lower performance).

    Both integrated some extra logic to simplify design. This meant it had more pins, which led to it being the first Intel CPU in the family delivered in a PLCC package instead of a DIP.

    Perhaps concidentally, the Motorola 68K family also had a chip that didn't go far, the 68010 (and 68012 and 68008), which integrated some logic and implemented some changes necessary for full virtualization. Apollo used this chip and I think Groupe Bull. The major proponent of 68K (Apple) skipped it.

    --
    http://lkml.org/lkml/2005/8/20/95