Clevo + Coffee Lake: Status?

Discussion in 'Sager and Clevo' started by thegh0sts, Aug 12, 2017.

  1. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,483
    Messages:
    13,089
    Likes Received:
    15,170
    Trophy Points:
    931
    The test results to find IPC differences speak for themselves :)
    no ipc gain between 7700k and 8700k.jpg
    no ipc gain between 7700k and 8700k #1.jpg
    The 7700K score is within a couple of points of result compared to the 8700k when comparing IPC at 4.5ghz.

    There is no IPC gain. There may be other losses or gains, but IPC is zilch difference between Kabylake 4 core and Kabylake 6 core.
     
  2. Papusan

    Papusan BGA Filthy = That sucks!! STAHP! Dont buy FILTH...

    Reputations:
    6,041
    Messages:
    14,397
    Likes Received:
    19,068
    Trophy Points:
    931
    See. 4.5GHz in CBR-15 from NBC.net
    upload_2017-10-13_2-20-3.png


    7700K@4.5GHz
    [​IMG]

    Of course you can get different scores!! Different tasks in the background and different memory will make different scores. How difficult must it be?
     
    ole!!!, hmscott and Mr. Fox like this.
  3. ole!!!

    ole!!! Notebook Prophet

    Reputations:
    1,042
    Messages:
    4,010
    Likes Received:
    1,873
    Trophy Points:
    231
    a test done by people doing brute force is no different than Edison, which is why there are individuals who truely understand science and concept are called geniuses, people like Einstein and Tesla.

    those tests are simply general test and i dont expect them to fully understand cpu architecture or what resources software choose to use but even after explaining if you still dont get it then i can say no more.
     
    hmscott likes this.
  4. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,483
    Messages:
    13,089
    Likes Received:
    15,170
    Trophy Points:
    931
    But the same reviewers that got the wide variance of 8700k results got well grouped results for the 7700k CB test, that's the point - the 8700k tests are all at once badly done (top 4 scores), and have a wide variation when properly at stock speeds - assumed, there still might be bugaboo's to be found in BIOS configurations.

    One score isn't a valid comparison in this case, the graph of all the reviewers on 7700k and 8700k CB results is more telling. And a variance of 1200cb - 1400cb is too wide to count on individual scores. Your CPU may not do as well as the top score, and more likely fall into the bottom performing Max Turbo category.
    comparing the wide range of 8700k cb scores.jpg
     
  5. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,483
    Messages:
    13,089
    Likes Received:
    15,170
    Trophy Points:
    931
    I don't think we need to continue if you won't look at the graphs and see the issues related to the results. :)
     
  6. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,483
    Messages:
    13,089
    Likes Received:
    15,170
    Trophy Points:
    931
    comparing the wide range of 8700k cb scores.jpg
    By this collection of CB results from a wide range of reviewers, taking off the top 4 scores for being non-stock, the remaining scores say that for the large part - 45% -, you won't get advertised Max Turbo, so I guess it's a good thing Intel doesn't guarantee it.

    So, yeah, Intel won't sell a lot of them when people discover they aren't performing.

    At least this "production" run has these problems. Maybe the next production runs in a few months will be better?
     
  7. ole!!!

    ole!!! Notebook Prophet

    Reputations:
    1,042
    Messages:
    4,010
    Likes Received:
    1,873
    Trophy Points:
    231
    the idea is not about the graph and not about reviewers testing a "wide variety" of 8700k. all 8700k are designed the same way, difference would be silicon so if clocked the same, they yield no difference.

    the term in IPC, once again, should technically be including the entire CPU which includes extra cache and faster IMC. if a software using the extra cache/IMC then you will see it being much faster than 7700k at same frequency per core, just how it is.

    a person riding a consumer motorcycle at max speed would not mean if the consumer bike is replaced with a race bike. hardware is different, software needs to take advantage of that difference. in this case the software didn't take advantage of that difference thus no IPC gain. if it did, we will see IPC increase.

    a nice example would be samsung SSD and its TLC cache to boost computing experience. hence me and @tilleroftheearth dont really like sammy ssd that much when it comes to steady state or consistent workloads. MLC ssd ive no problem with tho.
     
  8. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,483
    Messages:
    13,089
    Likes Received:
    15,170
    Trophy Points:
    931
    I think you may have gotten the point of the graphs after all :)

    The 8700k silicon isn't performing, with about 45% of the CB scores being under the average of 1425cb - with a range of 1262cb - 1364cb.

    We still don't know for sure if the 1425cb is totally stock, as this may all be down to motherboard makers interpretation of Intel's specs for "stock" defaults.
     
  9. ole!!!

    ole!!! Notebook Prophet

    Reputations:
    1,042
    Messages:
    4,010
    Likes Received:
    1,873
    Trophy Points:
    231
    i dont need to look at it to understand the graph i mean most reviewers do graph the same way, but it looks like you dont understand what IPC really is tho. but i'll leave it at that.
     
    hmscott likes this.
  10. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    3,483
    Messages:
    13,089
    Likes Received:
    15,170
    Trophy Points:
    931
    Yeah, I don't like the results either, no IPC improvement demonstrates that this isn't really Coffee Lake architecture it's 6 core Kaby Lake, but we were already told that by Intel, the real Coffee Lake is out past 2H18 at this point, a date range which can change again.

    Architecture changes compared to Kaby Lake
    Coffee Lake features largely the same CPU core and performance per MHz as Skylake/Kaby Lake.[7][8]
    https://en.wikipedia.org/wiki/Coffee_Lake
     
Loading...

Share This Page