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

 


Forgot your password?
Close
typodupeerror
×
Intel Hardware Technology

Intel Unveils 10th Gen 'Comet Lake' CPUs, Pricing (pcgamer.com) 69

Posted by BeauHD from the new-and-shiny dept.
UnknowingFool writes: Intel released more information about their next generation CPUs, codenamed Comet Lake. Overall, CPUs will get more cores and threads and slight speed boosts. Price wise, Intel is cutting prices to be more competitive with AMD's Rzyen processors. Some of the downsides include requiring new socket (thus new MBs), LGA 1200 and lack of PCIE 4.0 compatibility. No specific benchmarks were released, however Intel claims to have the fastest gaming CPUs. "[T]he top Comet Lake chip is the same price as the top Coffee Lake at $488, and the cheapest Core i3 is $122," reports PC Gamer. They expect the release date to be sometime in May, though no official date has been confirmed.
This discussion has been archived. No new comments can be posted.

Intel Unveils 10th Gen 'Comet Lake' CPUs, Pricing More

Intel Unveils 10th Gen 'Comet Lake' CPUs, Pricing

Comments Filter:
  • mildly interesting release, a little perplexed that haven't gone PCIE 4.0

    • Re: (Score:3, Interesting)

      by account_deleted ( 4530225 )
      Comment removed based on user account deletion

      • How about NVME cards, fast RAIDs, having enough headroom for 10Gb Ethernet, and those fancy GPUs?

      • Only if you game. Some types of GPGPU applications are bottlenecked by the bandwidth available to move data onto the GPU.

      • Re: (Score:2)

        by AmiMoJo ( 196126 )

        Bandwidth demands are increasing. PCIe gen 3 x4 only provides about 3.5GB/sec (with overheads) for an NVMe SSD, for example. So either you dedicate more lanes or you move to gen 4.

        USB 3 is already able to saturate gen 3 x4 links and USB 4 will push that even further.

        There is also the connection between the CPU and the chipset to consider. The GPU is connected directly to the CPU but most of the other stuff is on a shared 8x bus that goes to the chipset which then has its own lanes for peripherals, so with g

        • I thought about cross-platform games too when I learned about the special SSD in the PS5. As I understand Xbox's X SSD will be slower so I guess cross-plaform games will go fo the least common denominator. That probably means games will require an SSD on PCs too.
          PS5 exclusives won't be constrained so they might offer something not seen on the other platforms

    • need more then 16 lanes + DMI.
      and in most boards the 16 is routed to gpu at x16 or x8 x8

    • Comment removed based on user account deletion
      • maybe, but this leaves them with another whole generation or maybe 2 with AMD having a clear advantage for anything that makes heavy use of the PCIE bus.

        • Re: (Score:2)

          by AmiMoJo ( 196126 )

          At the moment PCIe gen 4 can actually be a disadvantage in some situations. NVMe SSDs, for example. Manufacturers have tuned them to get really high numbers in simple read/write benchmarks but not in actual application performance so for now it can actually be an advantage to use a gen 3 SSD that is properly tuned.

          Of course this will be sorted out sooner rather than later but it's an interesting thing to note. First generation products for any new tech are often kinda bad and of course the manufacturers los

        • Re: (Score:3)

          by fazig ( 2909523 )
          I've got PCIe 4.0 on my Ryzen 9 3900X with an ASUS ROG Crosshair VIII Hero. For 99% of my uses so far it's not been more useful than PCIe 3.0. And from what I gather from other people on the Zen 2 platform they mostly don't benefit from it either with the exception of some fringe cases.
          All it really did was introducing an annoying actively cooled chipset on the mobo. And which the brilliant ASUS engineers placed exactly at a point where it pulls in the hot air from your typical open air cooled graphics car
          • Comment removed based on user account deletion

            • Re: (Score:2)

              by fazig ( 2909523 )
              A sidegrade might be possible for you.
              Before I went for the expensive ASUS I used the much more affordable Gigabyte X570 Aorus Pro.

              On that board the chipset fan is placed in a more sensible position so it doesn't get covered up as easily. As a result the fan can operate at lower RPM and stays rather quiet.

              Unfortunately within the first week of use the Gigabyte board got stuck while trying to POST the RAM on a reboot. I tried leaving it trying to POST over night, but nothing happened. The clear CMOS jum

    • The boards will support PCIe 4.0 once Rocket Lake comes out next year.

    • That would require an actual redesign. These are just old wine in new bottles, and unlike wine they're not improving with age.

  • The big news is the low end i3 (Score:5, Informative)

    by rsilvergun ( 571051 ) on Thursday April 30, 2020 @08:44PM (#60009530)
    and it's Ryzen 3300x equivalent, which are rumored to be on par with a i7-7700k. In terms of affordability that would be a level of progress we haven't seen since the the Athlon 64.
    • competition is a damn good thing YAHOO!

    • Re: (Score:2)

      by AmiMoJo ( 196126 )

      Combined with an NVMe SSD that is as fast as RAM from the Core 2 era and modern PCs are finally making big enough strides that it's worth upgrading.

    • Re: (Score:2)

      by tlhIngan ( 30335 )

      Intel could always lower the price of their chips. They just chose not to. Presumably because they don't want to crush AMD which is bad (both for the market and Intel).

      Bad for market for obvious reasons. Bad for Intel because doing so means "free" patent licenses will end (AMD and Intel have so many cross-licensed patents that it's basically impossible to do an x86 competitor), lots of added government scrutiny, etc.

      So now that AMD's got a valuable chip again, and they're making reasonable amounts of money,

  • A 14nm failure (Score:3)

    by dicobalt ( 1536225 ) on Thursday April 30, 2020 @08:57PM (#60009546)
    that doesn't even bother trying to compete on price.

    • Re: (Score:3)

      by AmiMoJo ( 196126 )

      Intel just can't help itself. 14nm might be okay if the price was reduced and they supported their sockets a bit longer. Their attitude seems to be that they are a premium brand like Apple and people will pay for their shit no matter how bad it, as long as their new laptop has an Intel Inside sticker.

      • Re: (Score:2)

        by gweihir ( 88907 )

        Top be fair, there are enough idiots that will buy Intel simply because they have no clue and "nobody was ever fired for buying Intel".

    • Re: (Score:2)

      by gweihir ( 88907 )

      There are still enough idiots that buy Intel because they do not understand that Intel ripped them off and dies nit have anything competitive anymore.

  • 14nm (Score:3)

    by phalse phace ( 454635 ) on Thursday April 30, 2020 @08:59PM (#60009548)

    WTF? Intel's still putting out 14nm chips? AMD's already on 7nm and will be moving to 5nm next year.

    • Isn't Apple (with TSMC) already at 5nm? I can't wait to see the future Apple-designed A14 MacBook.

      • Re: (Score:1)

        by Anonymous Coward
        almost, scheduled for late this year (maybe next year with covid). Samsung is at 3nm for next year. But really you can't compare any of these measurements. the 5nm chips at TSMC have higher density than the 3nm chips from samsung. likewise for AMD and Intel, AMD has a much lower transistor density per nm scale, so their 7nm is roughly equal to Intels 10nm process.

      • Isn't Apple (with TSMC) already at 5nm?

        I wouldn't get concerned about marketing numbers. 5nm is meaningless. You can actually look up the different processes on Wikipedia and see that TSMC's 7nm process was actually worse in some metrics and only equal in others to Intel's 10nm process. And I'm not sure what Samsung call 7nm but their transistor density is damn bad compared to the competition "7nm" and Intels 10nm.

        These sizes havne't made sense since FinFET technology came out with the 22nm process almost 10 years ago.

    • Re: (Score:1)

      by Anonymous Coward
      14nm is still fine, you can't directly compare those numbers between manufacturers, you need to look at transistor density. AMD still ahead but no where near the level that 7 vs 14 indicates. AMD 7nm is about intel 10nm.

      • Okay, Robert Holmes Swan.

        • Okay, Robert Holmes Swan.

          Instead of accusing someone for pretending to be an Intel shill you could just look up this stuff online. You may learn something about why "nm" has been nothing more than a marketing term since FinFETs were released (hint which 7nm process of TSMC's are you talking about? There are multiple with different dimensions).

          Instead you just look like a tool.

    • WTF? Intel's still putting out 14nm chips? AMD's already on 7nm and will be moving to 5nm next year.

      Don't harp on marketing. The "nm" designation has been meaningless marketing since FinFETs hit the market back in the 22nm movement. They aren't related to any physical dimension.

      Also AMD doesn't manufacture any chips. I think you're talking about TSMC or Global Foundries.

      • Itâ(TM)s not a total marketing gimmick as decreasing feature size increases transistor density. Now fanboys shouldnâ(TM)t use the numbers as absolute comparisons. âoeTSMC 7nm pwns Intel 10nm because itâ(TM)s 3nm smallerâ isnt a great comparison because both processes may be creating the same relative sizes but are measured differently. Itâ(TM)s marketing what each foundry calls their processes. What matters is that each generation squeezes more transistors onto the die.

        The rel

        • Re: (Score:2)

          by fazig ( 2909523 )
          The relevant point should be how their products perform in applications that matter to the user vs the price the user has to pay.
          Objectively AMD wins that comparison in most disciplines right now. But how much of that can actually be attributed to the manufacturing process compared to differences through architecture and pricing policies remains an open question.

  • Woof (Score:4, Insightful)

    by pezpunk ( 205653 ) on Thursday April 30, 2020 @09:31PM (#60009594) Homepage

    glad i'm not an Intel stockholder.

    • I was sad that the pandemic and market crash didn't touch AMD, I was hoping to buy cheap stock. Intel, who knows how a big a discount would be enough?

      • Re: (Score:2)

        by spth ( 5126797 )
        It did touch AMD: AMD went down from 54 € in mid-february to 34 € in mid-march.

    • Lol AMD's stock went down when it released the whole Ryzen and Threadripper line.

  • Intel finally managed to pull its head out of its ass about something.

    • Intel finally managed to pull its head out of its ass about something.

      By releasing 40 lanes at half the speed of AMDs...

      PCI-e 4.0 devices have already started hitting the market.

    • My understanding unfortunately is that the words âoeup toâ were used a lot during the official Intel briefing. âoeUp to 40 lanesâoe. âoeUp to 5.3 GHz brieflyâ. Etc. Until there are actual CPUs and MBs it seems like inflated marketing numbers.

  • Because Intel has be suspiciously quiet on the question. Previously, they got their speed advancement over AMD only by an excessively insecure design.

    • I got the impression that the speed boosts were somewhat conditional. For example the top end CPUs could ramp up to 5.3 GHz**.

      **In short bursts only, if CPU temperatures were under 70C, other conditions apply, etc

    • Are these still insecure?

      Because Intel has be suspiciously quiet on the question. Previously, they got their speed advancement over AMD only by an excessively insecure design.

      Yes and no. The underliying microarchitecture is still insecure. But every single fix/mitigation is already baked into the microcode/firmware/OS/Applications from the factory, which means that the performance you meassure is actually the performance you get.

      Having said that, I have to point out three issues:
      1.) As I said, the microarchitecture is still insecure, so there may be more performance-sapping vulnerabilities lurking on the shadows that we do not know about. When those are discovered, we will get f

  • TDP is rated at the base frequency, not even the boost frequency, which is enabled by default.

    125w chip? nope. so lame.

Slashdot Top Deals

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Close