Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Intel Businesses Hardware Technology

Intel Addresses CPU Shortage: 'Supply Is Undoubtedly Tight' (crn.com) 163

Intel interim CEO Bob Swan publicly addressed the company's CPU shortage issue for the first time since July, when he acknowledged that meeting additional demand would be Intel's "biggest challenge." From a report: In a message posted to Intel's website Friday, Swan said the "surprising return" to growth in the PC market "has put pressure on [the company's] factory network." He added, "We're prioritizing the production of Intel Xeon and Intel Core processors so that collectively we can serve the high-performance segments of the market. That said, supply is undoubtedly tight, particularly at the entry-level of the PC market."

Intel partners and at least one distributor previously told CRN they were seeing a shortage of Intel's current generation, 14-nanometer CPUs, most notably in lower-end client processors.

This discussion has been archived. No new comments can be posted.

Intel Addresses CPU Shortage: 'Supply Is Undoubtedly Tight'

Comments Filter:
  • by Anonymous Coward

    Whats TIGHT is the NSA's hand up Intel's ASS. Fucking spyware os in our chips.

    TIME TO DISENCORPORATE INTEL.

  • If only there was another processor you could purchase. A processor that is fast and affordable and competes directly with Intel Xeon and Core CPUs. Perhaps made by a company whose name rhymes with OMG...
    • by Anonymous Coward

      I wish there was but there isn't for workloads in the 1-6 thread range (for GIS) or with regards to single core performance. We have put some workstation purchases on hold for 6 months because Intel Workstation prices are 35+% right now. We tested Ryzen 2950X & 2990WX based systems and they were 15-25% slower than our previous generation Intel Xeon W 2105/W2155 series boxes

      What am I missing as someone who doesnt know hardware but understands software? Why is AMD so often trumpeted here but rarely seen i

      • Re:If only... (Score:5, Insightful)

        by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Friday September 28, 2018 @12:51PM (#57390384) Homepage Journal

        And why does AMD always perform great in benchmarks but badly in real world tests compared to Intel?

        That hasn't been my experience at all. The K6 blew the doors off the P2, clock for clock, when you compiled for it. Of course, no commercial software was, but if you run Linux you can compile most things yourself and reap the rewards. And the original Athlon likewise absolutely slaughtered the P3, clock for clock. The FP performance was hilariously superior. Today, AMD only outperforms Intel abusively dollar for dollar, and yeah if you need maximum single thread performance you have to go with Intel. That only matters to gamers though, since everything else that needs much performance is multithreaded now.

        • Re:If only... (Score:5, Informative)

          by edwdig ( 47888 ) on Friday September 28, 2018 @01:37PM (#57390714)

          Today, AMD only outperforms Intel abusively dollar for dollar, and yeah if you need maximum single thread performance you have to go with Intel. That only matters to gamers though, since everything else that needs much performance is multithreaded now.

          A lot of tasks are only semi-parallel though. Take building code - you can compile lots of files in parallel, but then you have to link the executable on a single thread. Even though the majority of the work is done across multiple cores, the single core performance can still make a noticeable difference.

          • Agreed. I would summarize the causes as:

            1. Memory bandwidth
            2. Single-threaded performance

            Turning off hyper-threading on the Intel chips and rebenchmarking should help show how much of a factor single threaded performance is having.

          • I'm fairly confident that gold linker and lld are multithreaded.
          • Take building code - you can compile lots of files in parallel, but then you have to link the executable on a single thread. Even though the majority of the work is done across multiple cores, the single core performance can still make a noticeable difference.

            Except your example is terrible, because linking is trivial compared to compilation, and happens rapidly enough. Any software project large enough to have a long linking phase tends to have libraries in its makeup (not just external ones, but internal ones) and the various libraries can be linked in parallel.

            • I've been working on some kernel networking stack performance improvements for work, so the first thing that popped into my head was Linux kernel compilation.
              Linking does feel like its an appreciable amount of time compared to the compilation phase.
              For most things, that obviously isn't the case.
            • by dryeo ( 100693 )

              Mozilla moved away from static libraries, replacing them with fake libraries, basically just a list of object files that usually would have made up the static library as well as disassembling outside static libraries into object files and they also concated 8-16 source files together before compiling. All to speed up linking xul, which was taking hours on OSX and a long time on other platforms.

        • That only matters to gamers though

          It matters to a lot of people, but gamers are definitely not in that group. And game benchmarks will happily show that AMD and Intel are neck in neck for the similar spec'd CPUs. The only real difference is the AMD CPU will run you $50 cheaper.

          • It matters to a lot of people, but gamers are definitely not in that group. And game benchmarks will happily show that AMD and Intel are neck in neck for the similar spec'd CPUs. The only real difference is the AMD CPU will run you $50 cheaper.

            Lots of games, notably ones which don't have a console port, aren't aggressively multithreaded. IME Intel-based systems deliver higher minimum frame rates. Early Athlon days aside, they tend to be able to do more fp math.

            With that said, I have only AMD and ARM processors because I think the value of AMD CPUs is much better than Intel.

            • Lots of games

              I think you meant to say "few". Gaming benchmarks across a very wide range of games pit (for example) the Ryzen 2700X against an i7 7700K (at a $30 premium) and an i7 8700K (at a $150 premium). On some games the i5 8600k comes into the mix too which you can get for a $20 discount if you feel like getting 25% less cores and significantly less mutlithreadding performance for your dollar (i.e. you only play the small subset of games that are only single threadded).

              The only way you get a cost benefit from Intel

        • You responded to his question about modern processors with shit that is almost 20 years old. That is really not useful.

      • Re: (Score:2, Interesting)

        by Anonymous Coward

        I wish there was but there isn't for workloads in the 1-6 thread range (for GIS) or with regards to single core performance. We have put some workstation purchases on hold for 6 months because Intel Workstation prices are 35+% right now. We tested Ryzen 2950X & 2990WX based systems and they were 15-25% slower than our previous generation Intel Xeon W 2105/W2155 series boxes

        What am I missing as someone who doesnt know hardware but understands software? Why is AMD so often trumpeted here but rarely seen in the real world? And why does AMD always perform great in benchmarks but badly in real world tests compared to Intel?

        i'm sure your workstations were fully patched with spectre and meltdown mitigations in software and in microcode when you compared them to the AMD offerings. from what i been told, most of that gap is non-existent in patched systems. meanin the only reason intel was faster was because they cheated, and when the holes, i mean cheats, were closed they perform very closely in real world conditions.

      • Re: (Score:2, Insightful)

        by Anonymous Coward

        Are you running Windows? The Ryzen you tested use NUMA (especially the 2990WX which is four-way asymmetrical!) and sometimes fair badly because Windows supports it badly.
        Maybe you used slow RAM, the AMDs perform better when you just ignore the official RAM speed and run faster RAM (the frequency of the RAM controller and the interconnects between packs of four core and between dies depend on it). Not sure if you can find fast ECC RAM though like DDR4 3200 or more. The effect is very big, there's major perfo

      • by lamer01 ( 1097759 ) on Friday September 28, 2018 @02:08PM (#57390934)
        From what I've seen, Intel cpus take a huge hit when those fixes are compiled in. From 8%-20% according to Phoronix. AMD cpus take a hit as well although much smaller. If you take into account those hits, AMD CPUs are faster even clock for clock.
      • > What am I missing as someone who doesnt know hardware but understands software? Why is AMD so often trumpeted here but rarely seen in the real world? And why does AMD always perform great in benchmarks but badly in real world tests compared to Intel?

        These are all fantastic questions! I would love to know the answers as well.

        If I was to wager a guess it would be: memory bandwidth

        If you take a look at the games where Intel beats AMD you'll almost see it comes down to memory bandwidth.

        Any chance you could

        • Minor corrections:

          "Games" should "gaming benchmarks"

          Underclocking the Intel CPUs is done so you can guage scalability.

          You should also try underclocking your RAM (again to understand scalability)

          Likewise also benchmark overclocked RAM.

          With the combined underclocked, normal, and overclocked combinations of CPU vs RAM you can map a 2D chart and you'll be in a better position to guage WHY Intel is faster then AMD.

          In my 30+ years of programming & building PCs these rules of thumb haven't changed:

          * AMD provid

  • by Zorro ( 15797 ) on Friday September 28, 2018 @12:30PM (#57390230)

    AMD

    • Re: (Score:2, Insightful)

      by Anonymous Coward

      Sure, now that all the good CPUs are gone, it is time for the second choice.

  • by amorsen ( 7485 ) <benny+slashdot@amorsen.dk> on Friday September 28, 2018 @12:31PM (#57390238)

    Are there any hard numbers showing that this is caused by increased demand rather than constrained supply?

    I expect that some of the previous-generation factories are in the process of being retooled for 10nm? Is that not how Intel does it? If it is, that would limit the supply of the 14nm chips without yet being able to make up the shortfall with 10nm chips.

    • Are there any hard numbers showing that this is caused by increased demand rather than constrained supply?

      I feel sure that Intel puts just as much effort into providing large supplies of entry-level processors with low margins as it does into top-end processors with very high margins.

    • by thegarbz ( 1787294 ) on Friday September 28, 2018 @12:52PM (#57390394)

      Are there any hard numbers showing that this is caused by increased demand rather than constrained supply?

      Did you miss the news that in Q2 PCs experienced the largest demand in over 6 years after several years of decline?

      Most of this is being driven by corporate Windows 10 adoption as the deadline for Windows 7 is getting closer. Companies don't upgraded OSes, they hand out new PCs.

      • "Companies don't upgraded OSes, they hand out new PCs"

        This might be what some small companies do but most large companies (who have much bigger PC fleets) will have their own SOE so they're not going to stick with the OS the hardware vendor puts on it anyway.

        • will have their own SOE so they're not going to stick with the OS the hardware vendor puts on it anyway.

          I said nothing about vendor OSes. I said they distribute new computers when they roll out new OSes, of course they use their own image and flavour. Currently with the rollout to Windows 10 it makes sense to distribute / upgrade computers in the process.

          *I work for a very large company currently doing such a rollout. Our slightly larger competitor is in the top 10 of fortune 500, they are doing it too, as is our next competitor further down the list. Maybe other industries are different. I only have the one

    • Last I heard Intel was having some serious problems with yields surrounding 10nm

    • I expect that some of the previous-generation factories are in the process of being retooled for 10nm?

      Intel's fab upgrade strategy is n-2. That is, for a new process node (n), they upgrade fabs 2 generations old. n-1 fabs (the current mode) are left alone as Intel needs that capacity in the present time.

      So Intel is currently in the process of converting some 22nm capacity to 10nm. 14nm tabs are not being converted (Oregon dev space aside).

  • by Archtech ( 159117 ) on Friday September 28, 2018 @12:36PM (#57390266)

    "... supply is undoubtedly tight, particularly at the entry-level of the PC market".

    Has anyone heard news of forthcoming Intel processors that have secure architectures and actually adhere to those architectures?

    • Has anyone heard news of forthcoming Intel processors that have secure architectures and actually adhere to those architectures?

      Does anyone except for VM / Cloud service companies care?

      • Does anyone except for VM / Cloud service companies care?

        Only people who care about security. If you're not among them, please post your public-facing IP here so someone can have a good time with your network.

        • Only people who care about security.

          If people care about security they would actually read and assess the impacts of security vulnerabilities. Other people just care about news headlines and feel good measures.

          I care about security deeply so I research big stories when they hit. It is precisely that reason why I don't give a shit about Spectre or Meltdown. I'd give you my public facing IP address, but you could figure that out quite easily if you put even a token amount of effort in (hint: domain name similar to username tied to fixed IP poin

    • Has anyone heard news of forthcoming Intel processors that have secure architectures and actually adhere to those architectures?

      Yes, and we've discussed them in prior discussions here on Slashdot, but I'm too lazy to go back and look that up for you. Try googling. IIRC it won't be before 2H next year, or possibly the next year.

  • So that people can stick to their old hardware for longer and give Intel more time to make new chips and get the bugs out of the smaller nodes. It's also time to give out some more x86 licenses so we don't just have AMD as a competitor. Cyrix/VIA need to come back.
    • Or just buy a fancy new AMD.

    • Cyrix/VIA need to come back.

      Why? Neither one was competent to make a high-performance processor, and the low end has moved on to ARM for the power savings.

  • I wonder how much of this supply crunch is suppliers building inventory into their supply chains to give a little room to maneuver in case tariffs or other trade barriers get put in place? I know being in the UK, I am building up more inventory than normal to hopefully give everything time to calm down if there is a cliff-edge brexit in six months time. Multiply this sort of behavior by all the businesses with international supply chains and you have shortages and an economic boom.

    It will be quite interesti

    • For Intel they don't make their CPUs in China from what I know. They make their NAND chips in China which would affect SSD prices. As far as I know China is big producer of silicon for solar panels by not high purity silicon used for chips.
  • by Luthair ( 847766 ) on Friday September 28, 2018 @12:48PM (#57390354)
    We make a fuckton more money on $10,000 Xeons than the $50 Pentium CPU.
  • by StormReaver ( 59959 ) on Friday September 28, 2018 @12:50PM (#57390372)

    It seems like Intel bought into the whole "Post PC" nonsense. It would be interesting to find out if AMD was similarly hoodwinked, or whether it has a ready supply of both low end and high end processors to fill the vacuum left by Intel's mismanagement.

    This is an opportunity for AMD to get much closer to Intel's magical 20% of the server market.

    As many of us have said for the last several years, desktop PC's aren't going anywhere anytime soon. Mobile devices will augment, not replace, the desktop PC market. It is one of the many things that Star Trek accurately predicted back in the 1960's.

    • Re: (Score:2, Interesting)

      AMD is coming out with the Athlon 200GE (Ryzen-based) processor [techspot.com] for $55 to round out the low-end. Intel still haven't come out with a processor to compete with AMD's 32-core/64-thread Threadripper 2 processor.

      I'm planning to switch from AMD to Intel for my next PC upgrade because the feature I want in a motherboard (dual ultra m.2 slots) are only available in an Intel motherboard. If an identical motherboard appears for AMD, I would go with Ryzen instead.

      • Comment removed based on user account deletion
      • by Anonymous Coward

        This dual M.2 slot feature is most useful on small form factors like ITX or STX. Even better on a laptop that won't fit 2.5" drives anymore.

        On AMD you'd go with a full ATX board or even a micro ATX that has two PCIe 16x slots, one of which works in 4x. Then you have a slot right there for a 4x M.2 SSD, using a small adapter board.
        If upgrading to a used i7 5960X, 6900K etc. you'd have no shortage of big boy PCIe slots to put SSDs in.

      • Comment removed based on user account deletion
  • Give them a break, it takes time to produce CPU's with that many bugs, backdoors and loopholes and yet still mostly function as a CPU.

"When the going gets tough, the tough get empirical." -- Jon Carroll

Working...