1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

News AMD unveils Ryzen processors, boasts of Intel-beating performance

Discussion in 'Article Discussion' started by Gareth Halfacree, 14 Dec 2016.

  1. Isitari

    Isitari Member

    Joined:
    6 May 2009
    Posts:
    343
    Likes Received:
    19
  2. Anfield

    Anfield Well-Known Member

    Joined:
    15 Jan 2010
    Posts:
    4,734
    Likes Received:
    333
  3. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858
  4. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858
    1600x bench leaked.

    http://wccftech.com/amd-ryzen-5-1600x-cpu-benchmark-leak/

    For comparison purposes, we used a Core i5-7600K running in our test rig and loaded the same CPU-z version (v1.78.1 x64). The quad core (non hyper-threaded) chip achieved a score of 2130 in single-threaded and 8206 in multi-threaded.

    The Ryzen 5 1600X had a score of 1888 in single-threaded performance bench and 12544 points in multi-threaded performance bench

    This is not a conclusive result but the multi-threaded performance is really good for the AMD part. This chip is expected to retail at $259 US which is $17 more than the $242 asking price of the Intel part.

    lmfao.
     
  5. rollo

    rollo Well-Known Member

    Joined:
    16 May 2008
    Posts:
    7,700
    Likes Received:
    99
    So hang on a minute 4 cores no hyper threading is only 4K behind 12 threads. I expected the multi core to be a lot further ahead.
     
  6. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858

    It's 30% faster threading and costs $17 more. And there's a very high chance the I5 is clocked much higher. Remember, the I7 5820k has crap "out of the box" clock speeds. Look at the single core score, that pretty much says it all.

    AMD said they would compete with Broadwell E. Not Skylake and not Kabylake.
     
  7. Anfield

    Anfield Well-Known Member

    Joined:
    15 Jan 2010
    Posts:
    4,734
    Likes Received:
    333
    The 4 core cpu gets roughly two thirds of the score of the 6 core cpu.

    Hyperthreading doesn't provide all too much of an advantage (varies depending on software used of course), look how close the 7700k is to the 7600k:

    https://www.bit-tech.net/hardware/cpus/2016/12/05/core-i7-7700k-performance-overclock-preview/1

    And for the premium Intel charges with the 7700k over the 7600k just for the privilege of hyperthreading you'll get another two physical cores thrown in on the AMD side.
     
  8. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858
    Just so that Rollo doesn't implode...

    The I5s always beat the high end CPUs when it comes to single threaded performance. The 7600k slaughters my 5820k for gaming. Take it to something that threads, though? yeah, bye bye.

    I knew AMD were going to be aggressive with pricing but s**t darn, a 5820k for I5 money. That's amazing.
     
  9. rollo

    rollo Well-Known Member

    Joined:
    16 May 2008
    Posts:
    7,700
    Likes Received:
    99
    I am not really fussed either way, not personally in the market for a cpu. By the time I am zen will likely be in its second or 3rd version.

    Just wondering why 8 extra threads does not deliver more than 4K score wise. Poor scaling? Poor coding? Cpuz usually just 100% your cores and threads as it's a good stability test.
     
  10. Gareth Halfacree

    Gareth Halfacree WIIGII! Staff Administrator Super Moderator Moderator

    Joined:
    4 Dec 2007
    Posts:
    12,142
    Likes Received:
    1,673
    Hyper-Threading is basically misdirection: you can run double the simultaneous threads only if they're not all doing the same thing (i.e. aren't both trying to access the same shared logic). Think of it like eating a really long sandwich: you and I can eat it at the same time if we start at opposite ends, halving the time it takes to eat a sandwich (and meeting in the middle for a Lady and the Tramp smooch); if you insist on starting at the same end as me, though, we're going to have to alternate bites and it'll take just as long as if I were eating it alone. That's Hyper-Threading.

    So, what you've got here is a four-core chip versus a six-core chip. Simple maths: the four-core chip scores 8206 in the multithreading test, which is 2051.5 points per core; the six-core chip scores 12544, which is 2090 per core. In other words, the two chips are largely on-par performance wise, with Hyper-Threading giving a minor boost to the AMD chip.

    EDIT:
    Here's a little demonstration, courtesy my AMD A10-5800K - which any critic will tell you isn't a real quad-core because it shares some of the logic elements between core pairs. Here I'm running the SysBench CPU benchmark with differing numbers of threads, which is the absolute worst-case scenario for any shared-logic/Hyper-Threading system 'cos all the threads are trying to do exactly the same thing at exactly the same time.

    [​IMG]

    You'll note the biggest improvement in performance comes as a result of going from one thread to two threads, as I've doubled the resources available: it's almost double (1.9695x) the performance. Going from two to three, though, is less impressive because the third thread is starved of resources: now I'm getting 2.8834x the performance, instead of the expected 3x. Going to four ends up being even less of a gain: rather than the 4x boost you might expect, I'm only getting 3.6788x the performance. Naturally, increasing the threads still further does nothing - in fact, going to six threads hurts performance (3.628x the performance of a single thread).
     
    Last edited: 17 Feb 2017
  11. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858
    Not sure and TBH I CBA revisiting my bios a few times to find out :D

    What I may do though is run it on my stock 5820k.

    Edit. Stock 5820k 2133mhz ram.

    [​IMG]

    Overclocked to 4.5ghz

    [​IMG]

    And...

    [​IMG]

    $17 more than an I5.. $17 more than an I5...
     
    Last edited: 17 Feb 2017
  12. Anfield

    Anfield Well-Known Member

    Joined:
    15 Jan 2010
    Posts:
    4,734
    Likes Received:
    333
    Scaling with hyper threading is poor.

    Just did a quick test on a 6700k at stock clocks:

    Without hyperthreading: 8362
    With hyperthreading: 9216

    (note, that is on an older version of cpu-z, so not necessarily comparable to the 5820k results posted by Vault-Tec).
     
  13. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858
    I'm not sure tbh. Your score seems perfectly believable really. Mind you, you could always download 1.74 and see if it changes anything.
     
  14. Anfield

    Anfield Well-Known Member

    Joined:
    15 Jan 2010
    Posts:
    4,734
    Likes Received:
    333
    Got an irrelevant difference of 9 points with up to date version, so scratch the not necessarily comparable part.
     
  15. Vault-Tec

    Vault-Tec Green Plastic Watering Can

    Joined:
    30 Aug 2015
    Posts:
    8,958
    Likes Received:
    858
    Yeah I kinda figured as much tbh. I couldn't see them radically changing it in a newer version.
     
  16. Journeyer

    Journeyer Well-Known Member

    Joined:
    31 Aug 2006
    Posts:
    3,033
    Likes Received:
    97
  17. Gareth Halfacree

    Gareth Halfacree WIIGII! Staff Administrator Super Moderator Moderator

    Joined:
    4 Dec 2007
    Posts:
    12,142
    Likes Received:
    1,673
    AMD's just confirmed it: XFR is exclusive to the X-suffix chips. Boom.

    Ending Credits, I nominate Yorkshire Air Ambulance as the charity to receive the bet's payout - and good on you for taking me up on the bet in the first place!
     
  18. Anfield

    Anfield Well-Known Member

    Joined:
    15 Jan 2010
    Posts:
    4,734
    Likes Received:
    333
    http://www.kitguru.net/components/c...t-ryzen-7-cpus-set-to-hit-intel-pricing-hard/
     
  19. Panomama

    Panomama I once signed up on uniform dating

    Joined:
    11 Jul 2009
    Posts:
    1,063
    Likes Received:
    24
  20. Dogbert666

    Dogbert666 *Fewer Staff Administrator

    Joined:
    17 Jan 2010
    Posts:
    1,592
    Likes Received:
    131
    Actually, XFR is going to be on all three Ryzen 7 chips. AMD is being a bit confusing about this, but this is straight from the engineering team.
     

Share This Page