Adler Lake Cores Stronger Than Expected?

Discussion in 'Hardware' started by Steve S, Feb 17, 2021.

  1. Steve S

    Steve S Pen Pro - Senior Member Super Moderator

    Messages:
    8,329
    Likes Received:
    3,841
    Trophy Points:
    331
  2. sonichedgehog360

    sonichedgehog360 AKA Hifihedgehog Senior Member

    Messages:
    2,289
    Likes Received:
    1,879
    Trophy Points:
    181
    It better be because Rocket Lake has fallen massively short and then some. By around that time, it will be going head to head with Zen 3+ and soon after Zen 4, which is rumored to offer a 29% IPC improvement.
     
    Last edited: Mar 7, 2021 at 2:08 AM
  3. sonichedgehog360

    sonichedgehog360 AKA Hifihedgehog Senior Member

    Messages:
    2,289
    Likes Received:
    1,879
    Trophy Points:
    181
    [​IMG]

    Link:
    https://mobile.twitter.com/9550pro/status/1368224604074827780/photo/1

    And there you have it: the entire Alder Lake mobile lineup has been leaked from top to bottom. As you can observe here, they are extremely bullish on the small cores. For these small cores to work right and play nice in tandem with the large ones, the Windows scheduler will need many further significant refinements than what we saw with Lakefield to properly load balance or play traffic cop with system resource requests.

    The reasons for this are twofold. One, legacy and current code just isn’t suddenly or automatically going to know on its own to smartly hop between the high performance cores and the low power cores. And two is developers aren’t about to recompile all their software either to work at peak speeds for an amalgamated hybrid design when they target a homogenous core structure. This is a huge risk with a large opportunity for failure seeing that Lakefield could not be made to execute remotely as efficiently as their cheaper and lower core count Core M variants.

    Therefore, they will need to finally get this heterogenous processing problem solved cracked the first time around—with no room for slip-ups or sliding roadmaps—meaning both the software and the operating system will positively need to synergi.2 or mesh well with this radically different microarchitecture to accomplish it or else suffer the dire consequences. It will be imperative for them to mobilize their vast army of software developers as well as get Microsoft completely on board to do the impossible or this Alder Lake, much like Lakefield, is going to end up in the dustbin of history’s failed experiments.
     
    Last edited: Mar 7, 2021 at 2:14 AM
  4. desertlap

    desertlap Scribbler - Standard Member Senior Member

    Messages:
    2,986
    Likes Received:
    3,805
    Trophy Points:
    181
    FWIW I'm seeing the start of some parallels between now and the days of AMD's Athlons. At the time they weer clobbering Pentiums in benchmarks and were gaining market share in desktops with the OEMS (not so much with laptops, but very different market then)

    Intel's response was Core I, which was arguably the biggest revision to intel's chips since multiple cores. And at the same time AMD stalled/stumbled in moving the Athlons forward. The genius of Core I most broadly was how scalable it was, running far cooler and more power efficient at the low end (so they could go in to laptops) and at the medium to high end, things like usable multithreading and most importantly, turbo speed ups.

    So that all being said, the computer market of 2021 is VERY different, and Intel is facing threats not just from AMD, but credible ones from the various ARM licensees.

    What is different this time, is I don't see a leader in Intel akin to Andy Grove, who truly embodied the innovate or die mindset. Instead I see them now more in the mode of protecting their markets.

    Don't get me wrong, Intel still has huge structural advantages in the corporate and OEM markets especially where the breadth of up and downline support absolutely dwarfs everyone else. But that ultimately will only carry them so far when some tests show a $1K M1 device significantly our performing a $3,5K Intel one.

    PS: I do think the real short term loser might be Qualcomm as it looks like other than a few niche devices, they look to be left behind.

    The X factor in all of this of course is Microsoft, and if they are willing and able to get behind one of these alternate chip architectures and TRULY make some optimizations to Windows. There are hints that at least they are thinking along those lines with Windows X, but we all know the recent history with MS follow through.
     
Loading...
Similar Threads - Adler Lake Cores
  1. NBraz
    Replies:
    0
    Views:
    379
  2. Steve S
    Replies:
    1
    Views:
    879

Share This Page