Announcement

Collapse
No announcement yet.

Intel Announces The Core i7 8086K That Tops Out At 5.0GHz; 28-Core Chip Coming This Year

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Originally posted by coder View Post
    Nooo... They've had 28-core CPUs since Q3 of last year:

    https://ark.intel.com/products/12049...Cache-2_50-GHz

    Note that:
    • it's also 14 nm
    • its base clock is 2.5 GHz - exactly half of what was demo'd; I'm reading that its all-core peak is 3.2 GHz.
    • TDP is listed as 205 W
    Power dissipation generally increases proportionally to the square of frequency. Right here, this should tell you a few things about their demo.


    Nooo... Some external refrigeration unit was spotted under the table. Confirmed here:



    Definitely not a desktop CPU. Not even workstation - it used their server platform and probably burned well in excess of 500 W, which is far beyond what any of their current CPUs are spec'd for, much less the 95 W ceiling for their mainstream desktops.


    Nooo... this was most likely just a publicity stunt to upstage AMD's new 32-core Threadripper, although they are said to be launching a new batch server CPUs (Cascade Lake?). Some of those will presumably be updates to their existing 28-core models.
    I'm surprised I had to go this deep into the thread to find this comment, was expecting to be the first thing railed upon this article. After all, it is in fact a 8180 Xeon with the unlocked multiplier that is being used.


    Originally posted by fuzz View Post
    "28 Cores of Bulls#!t - Intel's "5GHz" Parlor Trick"

    https://www.youtube.com/watch?v=tRH0-QwhvVQ
    I see you're also a follower of Tech Jesus.

    Comment


    • #32
      Originally posted by nitroflow View Post
      I see you're also a follower of Tech Jesus.
      It's difficult to find journalists who at least attempt to be unbiased.

      Comment


      • #33
        Originally posted by fuzz View Post

        It's difficult to find journalists who at least attempt to be unbiased.
        Verily. Fortunately there are still some.

        Comment


        • #34
          Originally posted by fuzz View Post
          It's difficult to find journalists who at least attempt to be unbiased.
          I don't know about that, but I certainly think Michael needs to cover this to atone for his lack of appropriate journalistic skepticism:

          Comment


          • #35
            I don't see anything wrong with that 28 core demo. Even if the cooling is silly, so be it. Even if it draws 500W the performance per watt seems good, but would be much better at 4.2 or 4.4 GHz.
            It is somewhat stupid to go that far but it's just uncommon on CPU. GPUs did go not that far from it in burning a ton of watts, see Radeon R9 390X and R9 Fury X. One of these went up to 375 watts!
            IBM had multi-chip modules with POWER CPU running somewhere between 4 and 5 GHz. I don't have a TDP figure for their current mainframe CPU : 696 square mm of CPU running at 5.2 GHz. I am sure these things are not cooled by a 2 centimeter high copper heatsink.

            If Intel manages to shave a few nanometers of latencies with the follow up CPU, Cascade Lake, that'll be more interesting still.

            Does it suck? I don't know. This doesn't seem to suck balls. I bet, few people need or want it.
            But if they do release a desktop version of this, less extreme but still requiring a more regular looking watercooling, well I don't see what's the problem, aside that it would be Intel-priced. Why not a 375 watt version, such that it matches the record of a component (graphics card, the whole thing not just the GPU, but it's for the sake of an argument)

            You could equally say the Xeon version is ridiculous and no one wants it. Probably, at $10K. No idea what they sell for though. You also have to know or remember that "Xeon Platinum" is a replacement for Itanium, and EX series / Xeon E7.

            Comment


            • #36
              Real men use liquid nitrogen.
              Test signature

              Comment


              • #37
                Originally posted by bridgman View Post
                Real men use liquid nitrogen.
                Doesn't that kill chips?

                Plus, it would be a bit more difficult for them to hide.

                Comment


                • #38
                  Originally posted by grok View Post
                  I don't see anything wrong with that 28 core demo. Even if the cooling is silly, so be it. Even if it draws 500W the performance per watt seems good, but would be much better at 4.2 or 4.4 GHz.
                  What's wrong with it is that they set it up as if it were a demo of an imminent product, targeted at high-end desktop users and gamers. And absolutely no mention of it being overclocked.

                  All of that was a lie. When demoing such an extreme exercise in overclocking, you don't "forget" to mention it. It should figure prominently into the presentation.

                  In fact, it seems that the CPU's power consumption might well have been north of 1 kW - in a socket rated for no more than 265 W.



                  Now, had they done that and made it more of a stunt and a spectacle, then the reception would be completely different. I, for one, have a lot of respect for what they accomplished. But when it goes towards serving decidedly underhanded purpose of undermining AMD's soon-to-follow press conference, it's hard to have any positive feelings towards it.

                  Comment


                  • #39
                    Maybe the Intel employees thought it was really obvious. Your points are taken but it was obvious also to people a bit in the know about hardware and reporting on hardware sites made it pretty clear. This thing has monster truck wheels and vertical exhaust pipes with fire coming out of the pipes.
                    But this shows that you NEED to state the obvious when the obvious isn't obvious to everyone. And in the end we can't rule out bad intent.

                    Comment


                    • #40
                      Originally posted by grok View Post
                      Maybe the Intel employees thought it was really obvious. Your points are taken but it was obvious also to people a bit in the know about hardware and reporting on hardware sites made it pretty clear. This thing has monster truck wheels and vertical exhaust pipes with fire coming out of the pipes.
                      The link I posted is what they saw after Intel got busted. In the actual presentation, everything was buttoned up, the chiller was carefully hidden underneath a table, and the piping to/from it was painted black. Sure, the case had a window, but it also had a black, unlit interior and black tubing. From the audience, it's not obvious there's anything unusual about it.

                      Originally posted by grok View Post
                      But this shows that you NEED to state the obvious when the obvious isn't obvious to everyone.
                      This is how I feel about tech journalists, such as phoronix . It should've been obvious to him that this was smoke and mirrors. He has first-hand experience with how power-hungry their HCC Skylake Xeons are and how much power is consumed by each incremental bit of overclocking. He should also know that Intel's 10 nm is badly lagging and they always lead off new process nodes with the smallest chips - not 28-core monsters. ...and yet, he parroted exactly what Intel said/strongly-implied and hasn't even bothered to post any sort of correction.

                      Originally posted by grok View Post
                      And in the end we can't rule out bad intent.
                      I think you can rule it in when:
                      Intel's Gregory Bryant began the demo by explaining that the company is working to bring the new chip to market in Q4 of this year, saying, "Without further ado, I would like to present to you a new 28-core processor."
                      And:
                      At the end of the demo, which focused heavily on the performance benefits of the 5GHz clock speed, Bryant said, "What’s amazing is that trade-off, this actually being a 5GHz in single-threaded performance frequency and not...having to sacrifice that for this kind of multi-threaded performance, so you've got kind of the best of both worlds. So, you guys want to see us productize that thing? Tell you what, we'll bring that product to market in Q4 this year, and you'll be able to get it."
                      Source: https://www.tomshardware.com/news/in...ghz,37244.html

                      In spite of what you say, a lot of people aren't equipped to know that this was BS. They count on good journalists to tell them when something is suspect. In light of that, here's a reminder of what we're talking about:
                      Intel Announces The Core i7 8086K That Tops Out At 5.0GHz; 28-Core Chip Coming This Year
                      Written by Michael Larabel in Intel on 5 June 2018 at 05:42 AM EDT.
                      ...
                      At Computex, Intel also demonstrated the first single-socket, 28-core desktop CPU. This 28-core CPU can also top out at 5GHz and is a desktop CPU. This new 28-core Intel CPU is said to be launching by year's end.
                      The only good thing I can say about this is that at least he didn't repeat the part about it running all cores @ 5 GHz, though I think that was most likely unintentional. Anyway, coverage like this is exactly what Intel was going for. And unless Intel is strongly held to account, I think we can expect to see more of the same from them and others.
                      Last edited by coder; 09 June 2018, 11:09 PM.

                      Comment

                      Working...
                      X