ClockTuner for Ryzen (CTR) - and introduction guide and download

Discussion in 'Processors and motherboards AMD' started by Hilbert Hagedoorn, Sep 29, 2020.

  1. Dave thomas

    Dave thomas Guest

    Messages:
    2
    Likes Received:
    0
    GPU:
    1080
    OK just starting out reading the guide, 2 initial questions:

    The Windows version: after installing updates as of today mine reads as build 19042.572
    The BIOS AGESA - mine reads as castlepeakPI-SP3r3-1.0.0.4


    are these 2 factors OK to continue please?
     
  2. mannix

    mannix Active Member

    Messages:
    72
    Likes Received:
    18
    GPU:
    MSI GTX1070 ArmorOC
    Hi Yuri,
    Indeed the double VID solves this specific problem but know I can see the VID just goes crazy high, up to 1.55v, every now and then.
    It's a different issue and doesn't seem to be easy to fix.... I'm now checking the VID voltage and disabling the OC mode if it's going over 1450mV or re-trying the whole process if it's not in a 25mV tolerance.
    But it's very annoying that for a small amount of time it runs so high.
     
  3. 1usmus

    1usmus AMD Memory Guru

    Messages:
    50
    Likes Received:
    98
    GPU:
    Nvidia GTX 1080Ti
    I will add an additional VID check before starting the stress test. At the moment, only the protection system monitors this problem and stops all processes.

    If you run CTR and there are no warnings in the log then everything is fine.
     
    mannix likes this.
  4. Dave thomas

    Dave thomas Guest

    Messages:
    2
    Likes Received:
    0
    GPU:
    1080

  5. shazzagraz

    shazzagraz Guest

    Messages:
    9
    Likes Received:
    2
    GPU:
    Sapphire RX 5700XT
    You just got your answer. 1usmus' post right above yours.
     
    MattyAu likes this.
  6. Intelitek

    Intelitek Guest

    Messages:
    3
    Likes Received:
    0
    GPU:
    Nvidia
    Hi guys, Update

    I kid you not – I have tried all possible options – took mobo off to a test bench disconnected everything. (R5-3600 + MSI B550m Mortar on BIOSv1.44 + 130W TDP tower CPU cooler)

    - Reset BIOS from mobo header

    - Removed battery and all power from mobo and left it overnight

    - Replaced RAM

    - Replaced VGA

    - QFlashed BIOS to all possible ver. from 1.0 to 1.44

    - Removed CPU heatsink, checked and re-seated CPU.

    Still stuck at CPU light on boot. Nothing until I got a new Ryzen 9, popped it in and voila! Works straight away. So my new R5 is now officially toasted. The strangest thing is I just had it for last 3 days +1 day as dead ;) – I had it on default setting, absolutely no OC then I came across CTR tuner and read all about it watched youtube vids about setting and prereqs followed everything to the last letters and kaboom… its most likely my bad luck or year 2020. Don’t take me otherwise – I don’t blame this app for anything - I take full responsibility for my loss. I am here just to share my experience and feedback for future app enhancement and awareness for all the beta testers out there that things can go south-bound.
     
  7. Intelitek

    Intelitek Guest

    Messages:
    3
    Likes Received:
    0
    GPU:
    Nvidia
    Hi Kalin


    With all due respect, I wasn’t expecting anything that extreme. It had totally taken me off guard. So no, I wasn’t monitoring anything. Comes to knowledge of overclocking – hence using an app that will do all the guesswork and heavy-lifting for us and sticking all values according to “best-practices” and “safe” default values. Not here to kick off any dust - just sharing what I am going through.
     
  8. daghtus

    daghtus Member

    Messages:
    25
    Likes Received:
    6
    GPU:
    RTX 2080S, 8 GB
    That's a real bummer. You must have had a lemon CPU. Give it a try and send it back for RMA. Of course, the CPU has been running stock at all times (not even Auto OC or PBO)... :rolleyes:
     
  9. seamen

    seamen Member

    Messages:
    25
    Likes Received:
    1
    GPU:
    RTX 2070S/ 8
    [QUOTE = "NightWind, post: 5846896, member: 283103"] Хм, мне правда жаль этого парня.
    В любом случае уже было много жалоб, и для простого диагностического прогона и / или некоторых тестов вы не должны повредить свое оборудование.
    Ну, по крайней мере, до сих пор никогда не было никаких проблем с ручным управлением или с ryzen master и более старыми вещами в прошлом.
    С ctr, к сожалению, у меня была своя доля аппаратных сбросов пару раз с выпуском beta6 (только).
    В некоторых случаях люди действительно должны быть более осведомлены о том, что они делают, но давайте не будем забывать, что приложения также должны быть безопасными, иначе какой смысл в их использовании. [/ QUOTE]
    this program is safe - it just resets data in case of errors, if someone has burned chip this is not a program problem
     
  10. mannix

    mannix Active Member

    Messages:
    72
    Likes Received:
    18
    GPU:
    MSI GTX1070 ArmorOC
    Hi Yuri,
    I'm pretty confident I've found a solution for the 1.55V issue.
    Would be nice if you test as well and let me know your results.

    The culprit is? The AMD Ryzen Master Monitoring SDK. Thank you AMD...

    The voltage goes crazy only if you query the SDK while applying a voltage/clock.

    I could reproduce in 30 seconds before.
    Now it's been 45 minutes my 3800x is switching profile every second under CPU-z stress test, no occurrences.

    So you need to apply a shared mutex between the set clock/vid methods and the monitoring.
    If there's no concurrency there's no problem.
    It's also critical the user does not run Ryzen Master together with CTR.
    But as of today doesn't really matter because CTR itself will interfere.

    Now... did you say this is the problem why AMD retired the RM SDK that could change clocks and voltages?
    Wow, either they have no idea how their code (which seems a can of garbage) is actually working or they just lied.

    Anyway, I'm very disappointed in AMD; their support for "enthusiasts" is just a marketing jig.

    Their developers zone is awful, is between non-existent to "Am I back in the '80s???".
    Whoever knows the nVidia developers zone knows what I'm talking about.
    The difference is between middle-age and the Enterprise bridge.

    Maybe it's time to get rid of the AMD RM SDK and try to do everything via direct calls.
    I thought it was the best and safest way to access the Ryzen processor but clearly it's the opposite.
     
    Last edited: Oct 28, 2020
    1usmus, jonske and daghtus like this.

  11. santidb91

    santidb91 Guest

    Messages:
    1
    Likes Received:
    0
    GPU:
    RTX 2060
    Hello Yuri and everyone, first of all great tool! and second, I have a 3950x paired with a tuf gaming x570 plus and 64Gb of kingston hyperx predator 3200 MHz. I tested the ram and it's stable @3133 1.355v cl16.
    According to CTR the cpu is a platinum saple, but no matter what I try wether is ocing or downvolting it, it ends up with an unstable profile.

    This is the result for downvolting:
    [​IMG]

    and as for overclocking I ended up with:
    [​IMG]

    Both of the above mentioned are unstable (run it with Enhance accuracy and without, same result), my question is if it is possible to change the testing methodology for it to achieve a stable oc/downvolt?

    This is what I'm using atm and it's been running fine so far:

    [​IMG]

    Cheers,

    Santiago.
     
  12. Karel

    Karel Guest

    Messages:
    2
    Likes Received:
    0
    GPU:
    RTX 2070 Super
    I have a small problem/bug. I put my pc into sleep mode, if I wake it up again, it's not automatically oc'd. I have to update and apply profile every time. Not a big deal, but it may be something worth looking into.
     
  13. Ignatuz

    Ignatuz Guest

    Messages:
    8
    Likes Received:
    0
    GPU:
    GTX 1070 EVGA FTW
    Hello ! i have a question. My microprossesor is a r5 3600, watercooling 120 and 570x asus prime get gold in crt and i leave the program doing a test for undervolt because i need to do a thing outside home but when i come back crt fail and said the voltage is to high and was on 1.55 and i need to restart the pc immediately. i think the alert was for and hour in 1.55. This can affect the processor forever? now im using it whit 1175mV CCX1 (116) 4250 and CCX2 (127) 4275. Greetings
     
  14. mannix

    mannix Active Member

    Messages:
    72
    Likes Received:
    18
    GPU:
    MSI GTX1070 ArmorOC
    See my post above, it's a bug that will be fixed in the next version.
    If your processor is still working you don't have nothing to worry about.
    CTR is stopping testing with Prime if overvoltage or overtemperature is detected, hence the message.
     
  15. Ignatuz

    Ignatuz Guest

    Messages:
    8
    Likes Received:
    0
    GPU:
    GTX 1070 EVGA FTW
    omg! Thanks! you save my day! Greetings
     

  16. mannix

    mannix Active Member

    Messages:
    72
    Likes Received:
    18
    GPU:
    MSI GTX1070 ArmorOC
    It's a known limitation
     
  17. 1usmus

    1usmus AMD Memory Guru

    Messages:
    50
    Likes Received:
    98
    GPU:
    Nvidia GTX 1080Ti
    You're absolutely right, it is most likely the cause of 1.55 and some BSOD. I had hoped that they would use a more secure way to read data from the SMU, but they don't use Mutex.
    This is really a disaster. I've already started developing CTR 2.0 and I'll be sure to take your remark into account in it. Thank you so much.
     
    daghtus and mannix like this.
  18. bas_vdl

    bas_vdl Guest

    Messages:
    2
    Likes Received:
    0
    GPU:
    GTX1660 Super OC 6G
    I ran all version of CTR, and all end with the recommendation 4400 @ 1250. But when I set these values for my Ryzen 5 3600 in the BIOS and run the prime95 Small FFTs test, 3 workers immediately stop.

    Then I used Ryzen Master to set the following values and run the FFTs test (short runs). Below the different settings and results...

    4400 @ 1250, 1300, 1350, 1375 (fails)
    4300 @ 1250 and 1300 (fails)
    4300 @ 1375 (good, manually stopped after test 2)
    4350 @ 1375 (good, manually stopped after test 2)

    Wondering which stress test CTR is running?
     
    Last edited: Oct 30, 2020
  19. daghtus

    daghtus Member

    Messages:
    25
    Likes Received:
    6
    GPU:
    RTX 2080S, 8 GB
    I remember it ran higher FFT (2048 or 4096) in its earlier release. Not sure what it is running now but I believe the reason is the small FFT is not a real-life workload but a synthetic one which is clearly an overkill. I am currently running my 3600XT at 4500 MHz @ 1330 mV which would certainly fail small FFT but I have no problems whatsoever in games or with Boinc crunching.
     
  20. bas_vdl

    bas_vdl Guest

    Messages:
    2
    Likes Received:
    0
    GPU:
    GTX1660 Super OC 6G
    Also noticed that GTA 5 is crashingafter 30-45min of gameplay on 4400 @ 1250. What test you would suggest to test for stability?
     

Share This Page