AMD Cinebench Benchmark Demo at CES 2019 Buries the Current Intel Lineup

Discussion in '[H]ard|OCP Front Page News' started by cageymaru, Jan 11, 2019.

  1. illli

    illli [H]ard|Gawd

    Messages:
    1,162
    Joined:
    Oct 26, 2005
    I'm hoping this year we'll also see some good micro atx AM4 boards. I have not been too impressed by what is out there now.
     
    motqalden and auntjemima like this.
  2. BloodyIron

    BloodyIron 2[H]4U

    Messages:
    3,230
    Joined:
    Jul 11, 2005
    /micdrop

     
  3. BloodyIron

    BloodyIron 2[H]4U

    Messages:
    3,230
    Joined:
    Jul 11, 2005
    Referencing previous generation (Ryzen 7 2700x) as evidence of the performance of the generation following it, is just ignorant and a fool's errand. I'm not even going to entertain this topic since it seems you don't understand how that's completely irrelevant.

     
    Darth Kyrie likes this.
  4. BloodyIron

    BloodyIron 2[H]4U

    Messages:
    3,230
    Joined:
    Jul 11, 2005
    You know what I like?

    Computers. They're pretty neat.

     
    juanrga, tayunz and horrorshow like this.
  5. BloodyIron

    BloodyIron 2[H]4U

    Messages:
    3,230
    Joined:
    Jul 11, 2005
    I wouldn't be surprised. Clearly if something can't stand up to scrutiny, the next logical play is irrational fear mongering. Or FUD, that's a nice play Microsoft has perfected over the years.

     
  6. freeloader1969

    freeloader1969 2[H]4U

    Messages:
    2,222
    Joined:
    Dec 13, 2003
    Yes you are correct. His slides say "TBA...To be announced" at CES.


    Exactly. We're all Nostradamus when it comes to future product launches. :)

    Are you still around?
     
    Pieter3dnow likes this.
  7. Meaker

    Meaker Official representative of powernotebooks.com!

    Messages:
    922
    Joined:
    Jan 10, 2004
    If they refresh the cores to go with the ddr5 transition they could also refresh this platform with the modular design by keeping this I/O chip.

    It gives a lot of flexibility and should allow them to keep the apu side bang up to date.

    AMD are still way smaller than Intel and even best case will be for a long time. This let's them be agile with the minimum resources.

    Intel will struggle at the high core counts on a monolithic design to keep up. Less so at low core counts though.
     
  8. horrorshow

    horrorshow [H]ardness Supreme

    Messages:
    5,624
    Joined:
    Dec 14, 2007
    This made me LOL
     
    R_Type and BloodyIron like this.
  9. R_Type

    R_Type Limp Gawd

    Messages:
    156
    Joined:
    Mar 11, 2018
    The thing is Intel 10nm is due to finally drop this year, which will put them back on top again. Not by that much though.
     
  10. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    Back on top.. You sure of that?
     
    Pieter3dnow and auntjemima like this.
  11. trparky

    trparky Gawd

    Messages:
    883
    Joined:
    Jul 23, 2009
    Intel will eventually have to take a page out of AMD's book when it comes to making newer chips with more cores. Why? Because it simply comes down to complexity and the fact that as you try to add more cores to a chip the harder it is to get a good enough batch of chips to sell. You can see this in how the 9900K is so damn expensive. I have no doubt that it's because Intel has to either throw a lot of imperfect chips away or end up disabling cores to make lesser i5s and even lesser i3s.
     
    Darth Kyrie and Mega6 like this.
  12. R_Type

    R_Type Limp Gawd

    Messages:
    156
    Joined:
    Mar 11, 2018
    To be 100% fair thats not very [H] of you
     
    auntjemima likes this.
  13. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    Intel already is copying chiplet as you speak.
     
  14. IdiotInCharge

    IdiotInCharge Not the Idiot YOU are Looking for

    Messages:
    7,923
    Joined:
    Jun 13, 2003
    They've been waiting for their 10nm process to mature for what, four years now? And AMD has just barely caught up to Skylake in that time?

    Intel isn't always the best bet, but they're very rarely the worst.
     
    juanrga likes this.
  15. Zarathustra[H]

    Zarathustra[H] Official Forum Curmudgeon

    Messages:
    26,525
    Joined:
    Oct 29, 2000
    The [H] is not a "gaming" site. It is a PC hardware enthusiast site. :p
     
    N4CR, Darth Kyrie, motqalden and 2 others like this.
  16. IdiotInCharge

    IdiotInCharge Not the Idiot YOU are Looking for

    Messages:
    7,923
    Joined:
    Jun 13, 2003
    They've all been working on it for years... and it's not at all that novel. See: cell phone SoCs.
     
    juanrga likes this.
  17. trparky

    trparky Gawd

    Messages:
    883
    Joined:
    Jul 23, 2009
    Yep, so with that being said Intel is going to end up dealing with the same latency issues that AMD is dealing with.
     
    Darth Kyrie likes this.
  18. auntjemima

    auntjemima Hand Jobs Legend

    Messages:
    4,134
    Joined:
    Mar 1, 2014
    Bingo.
     
  19. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    if the implementation is so bad, why would intel follow - as usual?
     
    GHRTW and Darth Kyrie like this.
  20. auntjemima

    auntjemima Hand Jobs Legend

    Messages:
    4,134
    Joined:
    Mar 1, 2014
    Silicon is at its limits. There isn't a better way to go
     
  21. trparky

    trparky Gawd

    Messages:
    883
    Joined:
    Jul 23, 2009
    This. As a system becomes more complex (ie. more cores) the more chances of things going wrong during the manufacturing of the chip. The days of monolithic chips are over, we need to transition to multi-module chips (chiplet-based packages) or the cost of making the chips of the future is going to be absolutely hideous.
     
    Darth Kyrie likes this.
  22. trparky

    trparky Gawd

    Messages:
    883
    Joined:
    Jul 23, 2009
    Proof of this is in the Xeon line. They aren't expensive just because Intel likes to charge so damn much, it's because chips with that many cores in a monolithic package is [H]ard as fuck to manufacture. I'd not be surprised to see a failure rate of more than 65% and that's being optimistic.
     
  23. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    no. the proof is in the Zen line. Nice try.
     
  24. auntjemima

    auntjemima Hand Jobs Legend

    Messages:
    4,134
    Joined:
    Mar 1, 2014
    I don't think you know what you quoted.
     
  25. trparky

    trparky Gawd

    Messages:
    883
    Joined:
    Jul 23, 2009
    How so? Are you an Intel fanboy?
     
  26. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    I didnt quote a thing and ur right! Monolithic my bad.
     
  27. MaZa

    MaZa 2[H]4U

    Messages:
    2,608
    Joined:
    Sep 21, 2008
    Ignoring the bickering in this thread, I managed to watch the video only now and ho lee fuk if what he is saying is true, meaning this was just a 65W TDP lower midrange non-X model then the X models are going to be fricking awesome for both productivity and gaming. And the high end models with more cores are going to be absolutely bonkers for productivity work. Poor man's Semi-Threadrippers for everyone!
     
    N4CR, Pieter3dnow, Reality and 3 others like this.
  28. c3k

    c3k [H]ard|Gawd

    Messages:
    1,955
    Joined:
    Sep 8, 2007
    ^^^

    I usually start at the end of these threads and read to the start. Reminds me of the movie "Momento". ;)
     
    juanrga, Darth Kyrie and auntjemima like this.
  29. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    I like your attitude
     
  30. horrorshow

    horrorshow [H]ardness Supreme

    Messages:
    5,624
    Joined:
    Dec 14, 2007
    Well said sir
     
  31. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    You're right monolithic is old school.
     
  32. trparky

    trparky Gawd

    Messages:
    883
    Joined:
    Jul 23, 2009
    Yep. If we want chips with more cores the industry is going to have to adopt a multi-package or chiplet design or you can kiss the idea of more cores goodbye.

    Either that or enjoy $1000 chips and I'm sure nobody is going to like that.
     
    Last edited: Jan 12, 2019 at 8:00 PM
    Darth Kyrie likes this.
  33. R_Type

    R_Type Limp Gawd

    Messages:
    156
    Joined:
    Mar 11, 2018
    Oh yeah. The shrink will be good for them but more importantly the arch is getting some substantial improvements to. It'll push them back ahead. Ryzen 2 won't be exactly bad then but it won't be the home run either.
     
  34. Mega6

    Mega6 [H]ard|Gawd

    Messages:
    1,301
    Joined:
    Aug 13, 2017
    its already back and forth. Sorry, i don't see it happening. Should it happen, it will be short.
     
  35. IdiotInCharge

    IdiotInCharge Not the Idiot YOU are Looking for

    Messages:
    7,923
    Joined:
    Jun 13, 2003
    AMD is going back and forth with Skylake. AMD's going to need more than the small tweaks from Zen through Zen+ to Zen 2 to take the lead.
     
  36. R_Type

    R_Type Limp Gawd

    Messages:
    156
    Joined:
    Mar 11, 2018
    No I see zen2 being the better of Skylake even if the latter was on 10nm. Its the improvements in Icelake that will make the difference.
     
  37. Kurlon

    Kurlon [H]Lite

    Messages:
    72
    Joined:
    Dec 1, 2004
    I thought Intel debut the idea with the original Pentium Pro in 1995? Two separate silicon cores (Cache and CPU), one package.
     
  38. deepinya

    deepinya Gawd

    Messages:
    584
    Joined:
    Feb 5, 2003
    His voice is like nails on a chalkboard. lol
     
  39. bGm0

    bGm0 n00bie

    Messages:
    6
    Joined:
    Saturday
    Just to substantiate this point for all members. Thanks for the mention i did no knew K.H Chia. But that made me read his twitter posts... pretty good btw.
    So i found this chiakokhua/status/1060526913536458753 dated earlier than the graph you posted as damaging to AdoredTV`s Jim.

    Its pretty clear K.H. Chia knew and follows Jim. In another tweet basically uses previous AdoredTV video as source of new educated guesses.
    At least , the mentioned graph can be collaborative or inspired by...

    Bye the way, K.H. Chia has a whole thread on latency between chip-lets and IO die.
    The summary is this quote "The idea that Rome will experience a huge increase in latency just because the memory controller is moved to the I/O die needs to be debunked."
     
    Last edited: Jan 12, 2019 at 10:02 PM
    GHRTW, Darth Kyrie and schmide like this.
  40. DeeJayBump

    DeeJayBump Limp Gawd

    Messages:
    263
    Joined:
    Jun 18, 2002
    And Intel's 10nm was due [according to Intel] to drop in 2018, and in 2017, and in 2016, and originally in what 2015?

    At this point, what credence can Intel's claims regarding when either credible 10nm products or 10nm products that can combat Ryzen will release be given?

     
    Last edited: Jan 12, 2019 at 9:46 PM
    Pieter3dnow, GHRTW and Darth Kyrie like this.