Ryzen: Strictly technical

Discussion in 'CPUs and Overclocking' started by The Stilt, Mar 2, 2017.

  1. Ajay

    Ajay Platinum Member

    Joined:
    Jan 8, 2001
    Messages:
    2,848
    Likes Received:
    63
    Yep. OT: This will be very important for Raven Ridge. AMD/GLF need better than a 4 GHz 'Turbo' clock. I hope they can work it out.
     
  2. The Stilt

    The Stilt Golden Member

    Joined:
    Dec 5, 2015
    Messages:
    1,198
    Likes Received:
    1,321
    Raven isn't necessarily made on 14nm LPP ;)
     
    .vodka, Drazick, CatMerc and 3 others like this.
  3. Ajay

    Ajay Platinum Member

    Joined:
    Jan 8, 2001
    Messages:
    2,848
    Likes Received:
    63
    :eek:. I'm pretty sure you won't tell, but, what node might it be implemented on?

    Edit:

    Hmm, hadn't read about 14nm LPU - I've been slacking...
    10% higher perf @ iso power (supposedly).
     
    #128 Ajay, Mar 3, 2017
    Last edited: Mar 3, 2017
  4. imported_jjj

    imported_jjj Senior member

    Joined:
    Feb 14, 2009
    Messages:
    447
    Likes Received:
    278
    Raven Ridge is notebook first and i don't see why they need high clocks in desktop as it would be aimed at folks that don't buy discrete GPUs or do demanding things on it. Summit Ridge is the workhorse , RR is for less demanding users. Would make a difference for marketing and ASPs i guess.
     
  5. PPB

    PPB Golden Member

    Joined:
    Jul 5, 2013
    Messages:
    1,088
    Likes Received:
    134
    But arent other pstates affected by the nornal mode voltage offset, too? I thought all of them were affected by the undervolt offset of normal mode, and that caused more power consumption in all scenarios besides in P0

    Sent from my XT1040 using Tapatalk
     
  6. The Stilt

    The Stilt Golden Member

    Joined:
    Dec 5, 2015
    Messages:
    1,198
    Likes Received:
    1,321
    They are, but the other PStates are only used in "Normal-Mode", not in "OC-Mode".
    The lower PStates could specify 0.000V and it would make no difference, since they're never used in "OC-Mode".

    "OC-Mode" = P0 PState.
     
    Drazick likes this.
  7. PPB

    PPB Golden Member

    Joined:
    Jul 5, 2013
    Messages:
    1,088
    Likes Received:
    134
    Hmmm. That IMO seems like a poor decision, I mean, its almost like skylake no K overclocking where you are only left with p0 and idle and mixed load suffer power consumption suffer a lot from this.

    Sent from my XT1040 using Tapatalk
     
  8. dpnelson

    dpnelson Junior Member

    Joined:
    Mar 3, 2017
    Messages:
    1
    Likes Received:
    0
    My apologies if this has been asked and answered already, but does anybody know how the cores are ordered in Windows? On Linux, a 8C/16T device would have its cores ordered as ABCDEFGHabcdefgh. where "A" and "a" map to the same physical core. I haven't done affinity work on Windows, but my understanding is that Intel logical cores would be ordered AaBbCcDcEeFfGgHh. If Ryzen is representing them some other way (such as by core-complex, ABCDabcdEFGHefgh) that could cause all sorts of problems for applications setting affinity. If an application has one heavy thread (such as the "main" thread in many games), and locks it to one core while blocking other threads from using that core or its SMT-paired core, having the ordering wrong could lead to reduced performance of the main thread's core when other threads are accidentally sent to its SMT paired core.
     
  9. The Stilt

    The Stilt Golden Member

    Joined:
    Dec 5, 2015
    Messages:
    1,198
    Likes Received:
    1,321
    Odd cores are SMT in Windows.
     
    Drazick likes this.
  10. zir_blazer

    zir_blazer Senior member

    Joined:
    Jun 6, 2013
    Messages:
    852
    Likes Received:
    4
    Sorry The Stilt for hijacking a bit your interesing Thread, but for anyone that is into Passthrough as I do, I got the lspci output from Patrick from ServeTheHome from a Ryzen in an ASUS PRIME B350-PLUS. The PCI topology looks like this. I actually find it rather clean and flexible.
    I hear from this video that there seems to be poor isolation, and everything gets in the same IOMMU Group, or something like that (I don't see that they show log info or anything related to see how it actually looks). However, that means that they DID enable the IOMMU for the Linux Kernel (Else you won't get the IOMMU Groups constructs) and the Kernel didn't panicked when doing so. Since its an isolation issue, what seems to be broken is PCIe ACS, AMD-Vi/IOMMU itself works. However, with no further info is impossible to know what is broken.
    For reference, Skylake and Kaby Lake Chipsets also initially had ugly IOMMU Groups because the Chipset has a PCIe ACS related Errata, the ACS is found at an slight offset compared to the what the specifications says that it should be found at. Thus, for proper functionality, they require that you're using a Linux Kernel that had the fix to the quirk included so you get proper IOMMU Grouping. Since most people focuses on Ubuntu, chances are that if there was some last minute work or fixes for Ryzen that got include in the latest Linux Kernel, they are missing them. A bleeding edge distribution like Arch Linux, or a self compiled Linux Kernel from git, could be more interesing for Ryzen.


    A thing which extremely surprises me is Ryzen performance-per-Watt, is like AMD not only got competitive with Intel, but can actually beat it if you focus in that metric. I think that Ryzen weakness is the low Frequency ceiling, 4-4.2 GHz is not enough to put a good show against high Frequency Kaby Lakes in classical desktop workloads and gaming, but for Linux Server workloads it is AMAZINGLY IMPRESSIVE. Also, the binning that 1800X may require puts it extremely close to the "factory overclock" definition, taking away the performance-per-Watt and the fun of overclocking. Ryzen could put an excellent show if it stays at the 3.3 GHz Critical 1 Point, which is why the 1700 looks soo good.
    On Server workloads where performance-per-Watt is more important, is probable that Ryzen may be dramatically better. I think that ultimately, that is what will help AMD the most, especially from a profitability point of view, since AMD marketshare on Servers (Which got higher profit magins) was pretty much nonexistant since Sandy Bridge, and Ryzen could help get in with force. Desktop workloads doesn't showcase what it can truly do, it just shows its weakness. It could have made more sense if they focused on Server-first, as the original AMD K8 where Opterons came like 6 months before Athlons 64, but all the present issues shows instead that it would have been a product too inmature for that...
    I believe that Ryzen is "Bulldozer done right". I had the same expectatives from Ryzen that I had from Bulldozer (Reducing the ST gap with Intel to the "good enough" point, but dramatically better MT performance at similar price points), just that this time AMD delivered.

    BTW, what are the chances that we may see a midterm Ryzen refresh? I remember the Phenom II C2 vs C3 and Piledriver FX 8320E (Which was a more optimized Stepping. I recall a Thread from The Stilt talking about it some years ago). If Ryzen 4C/6C parts gets an extra 300-500 MHz headroom, it will seriously threat mainstream Kaby Lakes in ST. But at that point, it may have to face Coffee Lake instead...
     
    lightmanek likes this.
  11. Asterox

    Asterox Member

    Joined:
    May 15, 2012
    Messages:
    52
    Likes Received:
    37
    850 points in Cinebench 15 at 30W therefore 8 active CPU cores, what is CPU operating frequency at 30W?:cool:
     
  12. The Stilt

    The Stilt Golden Member

    Joined:
    Dec 5, 2015
    Messages:
    1,198
    Likes Received:
    1,321
    ~1900MHz average, based on the score.
     
    T1beriu and Drazick like this.
  13. hojnikb

    hojnikb Senior member

    Joined:
    Sep 18, 2014
    Messages:
    430
    Likes Received:
    15
    Is it possible for mobo makers to implement this "cTDP" feature ?

    Also, if you "lock" tdp down, does single thread or lightly threaded performance also suffer that much or is it smart enough to boost to over 3Ghz ?
     
  14. Asterox

    Asterox Member

    Joined:
    May 15, 2012
    Messages:
    52
    Likes Received:
    37
    Thanks, realistically this is new 8 Core CPU for Sony PS5.:D
     
    Agent-47 likes this.
  15. Agent-47

    Agent-47 Member

    Joined:
    Jan 17, 2017
    Messages:
    191
    Likes Received:
    156
    screw consoles. imagine a 30w 8c16t laptop that out performs a "top of the line" 220W FX :cool:
     
  16. The Stilt

    The Stilt Golden Member

    Joined:
    Dec 5, 2015
    Messages:
    1,198
    Likes Received:
    1,321
    It would be very easy for the ODMs to implement, if they had the knowledge to program the SMU outside the standard AGESA functions. Also since the feature isn't officially supported in consumer SKUs, AMD might simply tell the ODMs not to implement it in their bioses. I'm not saying they would, but it is entirely possible.

    Regarding the ST: 1800X at default, with turbo & XFR enabled scores 162 in Cinebench 15. With the TDP (PPT) limited to 30W the score is 155.

    Also once I figure out the way to safely (without risking the stability) undervolt the CPU in "Normal-Mode", then the performance at lower TDPs will be even more impressive.
     
    T1beriu, inf64, Drazick and 4 others like this.
  17. zir_blazer

    zir_blazer Senior member

    Joined:
    Jun 6, 2013
    Messages:
    852
    Likes Received:
    4
    Absolutely looking forward to look that.
     
  18. lightmanek

    lightmanek Junior Member

    Joined:
    Feb 19, 2017
    Messages:
    9
    Likes Received:
    21
    Amazing work, as always! Thanks The Stilt!
     
  19. someEEguy

    someEEguy Member

    Joined:
    Jun 5, 2013
    Messages:
    38
    Likes Received:
    5
    It seems reviewers have had some issues with early Asus and to a lesser extent MSI boards.
    http://www.techspot.com/review/1345-amd-ryzen-7-1800x-1700x/

    AMD@Gamers Nexus:
    https://youtu.be/TBf0lwikXyU?t=405
     
  20. lopri

    lopri Elite Member

    Joined:
    Jul 27, 2002
    Messages:
    12,096
    Likes Received:
    121
    Thank you, The Stilt, for this thorough and enlightening work. I read a few other "high-profile" reviews that are mostly low-quality (yet high on drama/hyperbole and eager to judge), and none of them gave me the insights The Stilt provided in this focused review. I am still trying to digest information provided in this thread and will most likely have some questions later, but for now I wanted to show my gratitude for The Stilt's high quality data as well as the sensible presentation.
     
    lightmanek likes this.
  21. Valantar

    Valantar Golden Member

    Joined:
    Aug 26, 2014
    Messages:
    1,166
    Likes Received:
    252
    So you're saying that cTDP limits total power draw over time, but only marginally limits peak turbo speeds. That's very interesting. Might we see mobile SKUs with specs along the lines of Ivy Bridge/Haswell 17W CPUs, with sub-2GHz Base clocks and ~+50% boost clocks (just 2-4x,the cores)? That would be very interesting. I'd gladly see 4c8t chips moving into the 15-25W mobile space, although with an iGPU thrown into the mix you'd probably need another 10+W of thermal headroom.
     
  22. leoneazzurro

    leoneazzurro Member

    Joined:
    Jul 26, 2016
    Messages:
    112
    Likes Received:
    39
    Do you think we will see a Ryzen migrating to the LPU process or something similar next year (as 7nm will be probably a 2019/2020 thing)? Or even partially to TSMC?
     
  23. laamanaator

    laamanaator Member

    Joined:
    Jul 15, 2015
    Messages:
    65
    Likes Received:
    10
    Perhaps Raven is made on the 14nm FDX (Or whatever the IBM 14nm process was called)?
     
  24. majord

    majord Senior member

    Joined:
    Jul 26, 2015
    Messages:
    217
    Likes Received:
    175
    I have a query or 2!

    1. Re SMU data.. I notice HWinfo is reporting Core powers, package power. Is this accurate when in 'OC' mode, with custom Vcore? I ask because it seems to be reading lower than expected at higher vcores.

    2. CPU-NB voltage. Does this have a new name? are still listing it as CPU-NB., and does athis plane acctually supply the DF?
     
  25. Rngwn

    Rngwn Member

    Joined:
    Dec 17, 2015
    Messages:
    119
    Likes Received:
    12
    So, this confirms that Ryzen does really supports windows 7, even better so than windows 10. Oh the irony. :p


    Any idea when this thing is going to get fixed?
     
    PotatoWithEarsOnSide likes this.