New CPU-Z Upgrade Lowers Ryzen performance

Discussion in 'Frontpage news' started by Hilbert Hagedoorn, May 4, 2017.

  1. Herem

    Herem Active Member

    Messages:
    86
    Likes Received:
    34
    GPU:
    Nvidia 1080
    I think the biggest problem here was changing the benchmark scale at the same time as dramatically alerting the relative performance. Had the scale remained the same if would have been clear if the Intel results suddenly got better due to enhancing the code. As the scale changed it makes it look suspiciously like the AMD results have been hamstrung.

    If it is the case that AMD chips really does perform certain code sequences very efficiently, then it would have been better for the developer to publish this as something for other developers to exploit and enhance their code rather than disabling it.
     
  2. Reikoji

    Reikoji New Member

    Messages:
    1
    Likes Received:
    0
    GPU:
    R9 290x
    First I know all the scores across the board are 'lower', but only because the scores bars started to go off the charts. I get that. But I thought the new scores were very fishy from the get go. First I figured, OK some of these reference scores have to be overclocks because theres no way 3ghz single core is scoring this high compared to 4ghz single core (and that is probably true), but now I see that practically every score you see for AMD processors have plummeted far beyond what intel's has.

    IF this was only about how Ryzen worked with [insert algorithm or instruction here], then why on earth did every other AMD processor plummet so much lower than Intel processors did? There were already side-by side comparisons showing older processors and how much lower they dropped. At first it was maybe 10% slower, now suddenly its 50% slower? Really? Is someone going to now say AMD processor scores on CPU-Z since its benchmarks conception have all been inflated? For YEARS?!

    Sorry but thanks to this, CPU-Z should be considered an invalid metric for determining anything, both before and after 1.79's benchmark 'fix'. The developer of it shows obvious bias for intel. The 'got sent a box of 6950ks' suggestion is probably not too far from the truth on t his.
     
    Last edited: May 7, 2017
  3. Athlonite

    Athlonite Maha Guru

    Messages:
    1,316
    Likes Received:
    36
    GPU:
    Pulse RX5700 8GB
    That's pretty much what I saw with my FX8320 @3.8GHz
    1.78.3 vs 1.79

    [​IMG]

    [​IMG]
     
    Last edited: May 8, 2017

Share This Page