Intel Loses Steam Thread

Discussion in 'Graphics and Semiconductor Industry' started by Raqia, Jan 21, 2014.

Tags:
  1. silent_guy

    Veteran Subscriber

    Joined:
    Mar 7, 2006
    Messages:
    3,754
    Likes Received:
    1,379
    How about you start by pointing out those continued serious mistakes?
     
  2. homerdog

    homerdog donator of the year
    Legend Veteran Subscriber

    Joined:
    Jul 25, 2008
    Messages:
    6,179
    Likes Received:
    964
    Location:
    still camping with a mauler
    I think NVIDIA has been doing quite well lately. Was very worried about 3 years ago, but they seem to be going full steam ahead these days.
     
  3. Sxotty

    Veteran

    Joined:
    Dec 11, 2002
    Messages:
    4,923
    Likes Received:
    352
    Location:
    PA USA
    Serious mistakes can be things like eutectic material that breaks down, mobile chips that don't sell, slowing pc markets, maybe or maybe not missing out on consoles. Many times one doesn't know until afterward. It isn't as if most companies decide to do obviously stupid things. This is not the thread though.
     
  4. RecessionCone

    Regular Subscriber

    Joined:
    Feb 27, 2010
    Messages:
    499
    Likes Received:
    177

    I think AMD's recent results show that even winning all three consoles would not have generated enough profits to change NVIDIA's fortunes significantly.
     
  5. Alexko

    Veteran Subscriber

    Joined:
    Aug 31, 2009
    Messages:
    4,499
    Likes Received:
    915
    I don't know that failing to win the consoles was a mistake as much as the inevitable consequence of the lack of powerful CPU cores. In my opinion, NVIDIA's main mistake was to refuse to be bought by Intel and AMD. When memory stacking becomes mainstream in PCs, I'd expect low- to mid-range GPUs to suffer immensely, and not being able to sell APUs is going to hurt NVIDIA very severely.

    I think they show just the opposite: it's precisely what keeps AMD profitable these days. But we should move this to the NVIDIA thread.
     
  6. silent_guy

    Veteran Subscriber

    Joined:
    Mar 7, 2006
    Messages:
    3,754
    Likes Received:
    1,379
    The packaging issue was made (not detected) over 7 years ago? Ancient history. If not, then why not include the 3Dmark 2001 (or 2003?) cheating as well? And blaming Nvidia, a $5B/year revenue company, for the slowing down of a $200B+/year market is giving them just a little bit too much credit. Why not point out that they have been stable in revenue despite the major declines of the PC market? The mobile chips could have done better, of course. They are taking on a Apple, Samsung, Qualcomm, Mediatek, and a bunch of Asian bottom feeders (combined $1T market cap?) all at the same time. It's surprising that they are still there, to be honest, when others have given up.

    I'm not saying that they are flawless and don't make mistakes, and their future is far from certain, but I have hard time coming up with the serious mistakes that you seem to be seeing everywhere and they are a remarkably strong company despite the environment that they are playing in.
     
  7. UniversalTruth

    Veteran

    Joined:
    Sep 5, 2010
    Messages:
    1,747
    Likes Received:
    22
    They are not helping and their, obviously, intentional actionless behaviour (lack of innovations which actually cause the mobile market boom) is actually an obstacle for this market

    As a big player out there, their responsibility should be taken into account
     
  8. Sxotty

    Veteran

    Joined:
    Dec 11, 2002
    Messages:
    4,923
    Likes Received:
    352
    Location:
    PA USA
    Wrong thread, but once again I am not saying it is their fault the industry is slowing down, but if they don't successfully deal with it (by diversifying into mobile for example) then that is their fault. Going down with the ship or even treading water might be nice, but doesn't show what is necessary.
     
  9. silent_guy

    Veteran Subscriber

    Joined:
    Mar 7, 2006
    Messages:
    3,754
    Likes Received:
    1,379
    If only they had tried to get into mobile. :wink:
     
  10. Entropy

    Veteran

    Joined:
    Feb 8, 2002
    Messages:
    3,072
    Likes Received:
    1,033
    Just read that Samsung missed their sales target for portable computers during 2013, selling only 12 million as opposed to the 17 million they had projected. They have adjusted their 2014 forecast down to 7 million units.
    They are expected to cease production of portable Windows computers after 2014, focusing on tablets and Chromebooks.

    Digitimes were the bearers of the news.
     
  11. liolio

    liolio Aquoiboniste
    Legend

    Joined:
    Jun 28, 2005
    Messages:
    5,723
    Likes Received:
    193
    Location:
    Stateless
    Read the news yesterday night. OT- ish but it reinforces my belief that neither Sony or MSFT are safe in the living room.
    Phones refresh rate is slowing down, I expect tablets to follow suit within a couple of years at max.
    Bad stars alignment, living room will be the only place for OS sellers and hardware manufacturers alike to grow their reach.
    They may have no choice but to bring the fight in the living room.
     
  12. Entropy

    Veteran

    Joined:
    Feb 8, 2002
    Messages:
    3,072
    Likes Received:
    1,033
    And where can Intel find margins for their processors/process tech in the living room? Where, additionally, the rate of replacement of devices is very slow?

    They are currently loosing money trying to get a foothold in mobile where the volume of shipped devices is large, but the margins are very thin indeed for the SoCs. There are only two manufacturers in mobile that makes any profit at all, and that's Samsung and Apple. Why would they change to Intel as a supplier, when they already have in-house solutions or very inexpensive ARM alternatives?

    Intel in Mobile is doing what Microsoft did with Zune, Kin, Bing, and so on - trying to buy their way into what the execs perceive to be "the next big thing" without a clear idea of either what would make their offering compelling or how to turn their loss leader efforts around to create sufficient margins going forward to make the whole endeavor worthwhile in the long term.
    Even if Intel managed to buy themselves 10 or 20% of the mobile market what the hell good would it do them?

    What Intel needs to do is to find/build a new market which allows them to leverage their strengths. It is embarrassingly obvious that their leadership, like Microsofts, have failed to do so. Giant cash warchests put to no good use.
     
  13. liolio

    liolio Aquoiboniste
    Legend

    Joined:
    Jun 28, 2005
    Messages:
    5,723
    Likes Received:
    193
    Location:
    Stateless
    I was not speaking of Intel, more Samsung and all those hardware and software vendor that ride the Android wave. Intel they want to be present in the Android realm, it is their choice. Now I'm not sure they can avoid part of the current race to the bottom, cheap hardware is simply getting too good.
    Now I don't know which new market Intel could find. Imo they invest what seems to be a sane amount of money in GPU tech, they are willing to cut their margin (with the atom line), I find it weird that securing a console deal was not on of their priority. It is not much in volume but would have kept their fab busy and buy them some prestige.

    I can't think of a new market with high margin, I think you missread my post, I don't think that smart tv and matching are to be a new eldorado, ulitmately it is commodity with volume and low margin. But I think there is some growth to extract here at a time when phones sales are slowing and I suspect tablet will follow suit soon. Gaming is relevant to the living hence my comment about Sony and MSFT not being "safe" in the living room.
    I'm not a big believer in the "internet side of thing" I simply know nobody that cares, the same applies to wearable. A niche for tech lovers at best.
    Now what Intel will is another topic, fight to have their SoC here? Stick to higher margin products? I don't know and my reaction was foremost about samsung results.
     
  14. Andrew Lauritzen

    Moderator Veteran

    Joined:
    May 21, 2004
    Messages:
    2,526
    Likes Received:
    454
    Location:
    British Columbia, Canada
    How much would that even affect it? Even if Intel were in both new consoles how big a %ge of their ~300mil per year desktop/laptop CPUs (not even counting server, workstation, mobile) would adding consoles constitute? 2-3% at best?

    I can buy the prestige argument (effectively a "halo" type argument) but the consoles are kind of small beans in terms of hardware units and revenue...
     
  15. liolio

    liolio Aquoiboniste
    Legend

    Joined:
    Jun 28, 2005
    Messages:
    5,723
    Likes Received:
    193
    Location:
    Stateless
    I definitely agree, it is not important still it seems you guys are working really hard (on GPU) and it would have been a nice showcase.
    Now I've no worries for Intel, the economy still sucks, there is stronger and stronger move toward commodity hardware but I'm having a hard time figuring out what could go wrong for the company:
    competitive low power SoC, the best high performance CPU cores around, great GPU, another take on parallel processors through larrabee's heirs, best process around.
    Lots of options.
     
  16. Andrew Lauritzen

    Moderator Veteran

    Joined:
    May 21, 2004
    Messages:
    2,526
    Likes Received:
    454
    Location:
    British Columbia, Canada
    Yeah. TBH Larrabee as a console would have been fun too, but the PC-like setup is probably better for developers overall :)

    I'm just an engineer so I'll repeat my signature in that I know nothing special here and this is just my opinion:

    I'm not really concerned in terms of tech. It's hard to see a future where - for instance - Apple does great and Intel goes out of business. The industry-wide "risk" is more the high end/premium stuff effectively disappearing entirely, or at least the economics of making it not working out. i.e. ending up in a world where $50-100 mediatek phones and tablets win. As long as there is enough demand for higher end stuff, I don't think Intel will have a problem providing competitive or leading solutions.

    As a slight digression... in terms of enthusiast PC gamers the question is a bit more interesting. Clearly for a few years at least we'll get to continue to ride the waves but ultimately if push comes to shove with "mobile first", it's not clear that high end hardware costs enough to support the ever-more-expensive R&D. So either enthusiast gamers get lumped into the "professional" category and those sorts of prices or they are just get a scaled-up mobile SKU that may or may not be particularly efficient. To be clear, as an enthusiast gamer I don't think it's all doom and gloom (quite the contrary, gaming on PCs has never been better!) but with everyone being forced to design primarily for mobile, it's going to have an effect.
     
  17. Raqia

    Regular

    Joined:
    Oct 31, 2003
    Messages:
    508
    Likes Received:
    18
    It shouldn't cost Intel too much to re-purpose a server board as a Skulltrail-like high end enthusiast platform so I think that market is alive and well so long as they make x86 servers. They can do their usual fusing off of professional features like virtualization etc. and milk their server hardware for extra revenues with lower budget enthusiasts in the near term.

    The long term question is whether enthusiasts will want to build a rig at all in the future. I think APUs should be much faster than the CPU+GPU across narrow PCI-E bridge paradigm that we use today once they're yoked to sufficiently fast and wide memory. Will there be a point to building your own system when there's so much integration on die, with perhaps a BGA interface for the APU and soldered GDDRX memory and precious little left to customize for the enthusiast? Perhaps there'll only be a comparatively small boost over mail ordering a Dell. Server like configurations with 20+ lower clocked cores and off die GPUs might be increasingly poorly suited for actual gaming in the long run.

    We might see starting with Broadwell, and its Gen 8 graphics.
     
  18. Andrew Lauritzen

    Moderator Veteran

    Joined:
    May 21, 2004
    Messages:
    2,526
    Likes Received:
    454
    Location:
    British Columbia, Canada
    Right, that's what I mean - i.e. the IVB-E type strategy. But while they have actually gotten somewhat more affordable in the last few years, these are still expensive chips compared to the $100-200 CPUs that I imagine most people buy.

    I agree that whether or not people will still build desktops is a little up in the air. That said, there's some forces that will still push the benefits of high-power GPUs (200W+) in the near future such as 4k and VR, the latter of which would be a mistake to underestimate I think. The other interesting tidbit is that while the overall PC market has been in decline, the gaming/enthusiast hardware market has actually continued to grow. Guess we'll see how it all works out in the end, but I do think there are a lot of question marks :)
     
  19. Blazkowicz

    Legend Veteran

    Joined:
    Dec 24, 2004
    Messages:
    5,607
    Likes Received:
    256
    Isn't Intel getting into microservers too?, they have an 8-core Atom for that. Some Silvermont product will be sold as "Pentium", too.
    This discussion so much feels like "big" traditional cores will be used for higher end markets only, sort of like POWER, Itanium and Sparc (or the Alpha, MIPS, Sparc etc. of the 1990s)

    Still, big cores are alive and there with Broadwell, * Lake, Steamroller, Excavator etc. and I guess we like the single-threaded performance fine, thanks.
    Will chips go asymetric eventually?, e.g. something like four "i7 cores" plus twelve "Atom cores" on one chip, with the assorted glue.

    I have a practical question too : maybe the socket will die, and we maybe shouldn't worry too much. Even now there's some good value on socket-less desktop boards if you're just needing a computer : AMD E-350, Celeron 1007. But can we keep DIMM slots on future offerings, thanks :razz:
    Feel free having like 1 to 4GB on package (if that becomes economically workable) but let me add 16 to 128GB (or more) "slow memory" on the side.
     
  20. Entropy

    Veteran

    Joined:
    Feb 8, 2002
    Messages:
    3,072
    Likes Received:
    1,033
    That's one way of looking at it I guess, although I see it more as micro servers encroaching on Intels turf, and Intel trying to stem the tide by offering similar products. It's a choice between two evils for Intel. Either they do nothing as ARM based platform start eroding their market from below, or they compete with similar products, thus helping legitimize and develop a market niche they would really prefer didn't exist at all, and where their margins are much thinner.

    The same goes for the entire mobile segment, where Intel actually sold off their StrongARM branch to Marvell, proceeded to cripple the netbook segment together with Microsoft, only to loose that entire market segment and more to tablets/convertibles/phones using ARM CPUs. They are trying to subsidize their way into at least tablets, but so far they have been unable to gain any kind of foothold in the market, and again it is clear that they face a choice between evils - either get eaten slowly from below, or take active part in building an ecosystem that erodes their bread and butter business.

    At best, doing battle there is a defensive action. There is no way Intel can hope to gain the market dominance that has allowed them the margins of the Wintel hegemony. Incidentally, today digitimes posted a rumor about Intel withdrawing from phones entirely. link
    It's a rumor only, but it does make sense.

    It's pointless for a company of Intels size to engage in rounding error businesses. For comparison, AMD enjoyed the launch of both the XBoxOne and the PS4, plus the holiday rush for the WiiU, and the total of their "embedded and semi-custom businesses" was just over a couple of percent of Intels revenue. And the margins of those embedded chips is comparatively low. Should Intel have tied up engineering resources to develop dedicated silicon for Nintendo/Sony/MS that they then could only sell at even thinner margins than AMD gets in order to secure the contracts? The ROI would be dismal, particularly by Intels standards, and any success of the consoles reduces their PC processor revenues. A lousy proposition.

    Intel needs to find a number of leads for the future, that preferably ties into the core competences of the company, and, even better, might reinforce each other. Then put their muscle behind building these new markets with the hope of three or four growing into something substantial. But finding those leads and how to nurture them is easier said than done. It might not even be possible.

    Incidentally, Intel Capital, their financial arm, has for example invested more than two billion USD in asian start-up companies. Apart from their golden egg laying geese in the CPU market, it may well be that Intels financial investments bring better prospects for the future than their manufacturing ones.
     
Loading...

Share This Page

  • About Us

    Beyond3D has been around for over a decade and prides itself on being the best place on the web for in-depth, technically-driven discussion and analysis of 3D graphics hardware. If you love pixels and transistors, you've come to the right place!

    Beyond3D is proudly published by GPU Tools Ltd.
Loading...