Results 1 to 10 of 10

Thread: Ryzen 5000 P0 OC Troubles...

  1. #1
    Xtreme Enthusiast
    Join Date
    Feb 2008
    Posts
    832

    Ryzen 5000 P0 OC Troubles...

    I've been having some troubles getting this board to behave as I would expect. I'd like a power limit, while having an all core overclock, and still have the voltage/frequency dynamically adjust to load. Any time I change anything to do with the frequency/voltage behavior, the board kills the power/current/temp limit. The chip will hit 105C+, take 250W in P95, and keep climbing. Any attempt at a P state OC, still results in no apparent thermal limit/throttling, no power limit, and it never throttles from P0.

    Is this really how it is, is it this board, or am I doing it entirely wrong? I don't care for stock behavior with PBO, the voltage/temps are sky high at idle, and the thing never appears to throttle. It DOES have a power limit though in this state.

    Asus Pro WS X570-Ace w/5900X

  2. #2
    Xtreme Cruncher
    Join Date
    Nov 2008
    Location
    NE Ohio, USA
    Posts
    1,591
    What CPU cooler you using? Either it isn't up to the task for the 5900X or it's not seated well on the chip.

    Don't set all an core ratio. Just use the built in OC behavior but lower the volts or TDP/power limit to get the temps under control and where you want.
    24/7 Cruncher #1
    Crosshair VII Hero, Ryzen 3900X, 4.0 GHz @ 1.225v, Fractal Design Celcius S36 AIO, 4x8GB GSKILL 3600MHz C15, 2x MSI 1660ti Armor OC
    Samsung 970 EVO 250GB NVMe, 2x10TB WD Red RAID1, BR Burner, Win 10 Pro, Enthoo Luxe TG, SeaSonic Focus+ Platinum 850W

    24/7 Cruncher #2
    GA-P67A-UD4-B3 BIOS F8 mod, 2600k (L051B138) @ 4.5 GHz, 1.260v full load, Arctic Liquid 120, (Boots Win7 @ 5.6 GHz per Massman binning)
    Samsung Green 2x4GB @2133 C10, 2x MSI 1660ti, Vertex 4 128GB, 2x3TB WD Red RAID1, BR Burner, Win7 Ult x64, Enthoo Pro, HX750

    24/7 Cruncher #3 ... Crucial M225 64GB SSD Donated to Endurance Testing (Died at 968 TB of writes...no that is not a typo!)
    GA-EP45T-UD3LR BIOS F10 modded, Q6600 G0 VID 1.212 (L731B536), 3.6 GHz 9x400 @ 1.312v full load, Zerotherm Zen FZ120
    OCZ 2x2GB DDR3-1600MHz C7, Gigabyte 7950 @1200/1250, Crucial MX100 128GB, 2x1TB WD Red RAID1, Win7 Ent x64, Centurion 590, XFX PRO650W

    Music System
    SB Server->SB Touch w/Android Tablet as a remote->S/PDIF->Denon AVR-X3300W->JBL Studio Series Floorstanding Speakers


  3. #3
    Xtreme Enthusiast
    Join Date
    Feb 2008
    Posts
    832
    It has an NH-D14 on it now, but I had a D15, and a H105, more or less the same temps. Temps are fine when power consumption is below ~150-175W, when its well over 200W on air it falls apart. This isn't the problem, the problem is the board appears to be ignoring the set current/power limits. Under stock operation they show up in HWinfo, but when I touch most anything related to the ratio/voltage/VID/P0, the limits are removed and it appears to let the chip burn down.

    If I change the power limit, all power and thermal limits are removed. If I adjust voltage, all limits are removed. If I do mostly anything, it will consume huge power in P95, not throttle, and maintain full boost clock 24/7 at ~1.48-1.52Vcore. The only way to make it not melt in P95 is to lock Vcore to ~1.125V or lower, and 42 ratio or lower. Even touching the curve optimizer removes all limits and breaks stock boost behavior.

    I sent a message to ASUS, maybe it's not supposed to do this and they're not aware of it. I already replaced the board due to a corrupt BIOS, but its the same behavior still.
    Last edited by ZOMGVTEK; 12-08-2020 at 09:09 AM.

  4. #4
    Brilliant Idiot
    Join Date
    Jan 2005
    Location
    Hell on Earth
    Posts
    11,112
    The tldr on anything not 5600x is don't manual OC it. your not going to be able to tame temps. I'll give you an example. My 5800x can do 4.5 fixed at 1.2v prime stable 81c peak. any attempt at 4.6 and 1.225v= over temp shutdown. Observing behavior at stock with PBO enabled prime runs between 4.4 and 4.8 depending on iteration. conclusion the only way you can beat amd @ overclocking is if you had 3 joysticks on board allowing you to control on the fly voltage clock speed and power limits and even then all you would end up with is 3 broken joysticks.

    This becomes even more of an issue beyond 5800x. This is with a corsair h110i which beats my nh-d15 and dark rock pro 4, I doubt even custom water would make much difference as i shoved radiator outside window and netted same outcome.

    CHVIII hero wifi. theres a tool that is sort of useful to.....to put things in perspective the 5950x gets a max clock load rating of 4000 mhz. which puts the typical 5900x somewhere in the middle and your's is doing 4.2

    Also take note that there is a rather significant difference in sil quality between ccd 0 and ccd 1 and ccd 1 dictates what your max stable load clocks are going to be as its much weaker than ccd 0 anywhere from 150 mhz to 300mhz that i have seen so far. You can find out how bad or good your ccd's are by running tool then disabling ccd 1. the first sil rating will go by lowest common denominator aka ccd 1 and then disable in bios to get the sil rating of ccd 0.

    So far i've seen sil ratings as low as 112 and as high as 128 with 117-122 being the avg.

    example






    Last edited by chew*; 12-08-2020 at 10:44 AM.
    heatware chew*
    I've got no strings to hold me down.
    To make me fret, or make me frown.
    I had strings but now I'm free.
    There are no strings on me

  5. #5
    Xtreme Enthusiast
    Join Date
    Feb 2008
    Posts
    832
    Dang. Thanks for the info. I had it running a manual OC since something odd is going on and it never goes below full boost at idle. I was getting around 1.5V claimed Vcore at idle, 0% load, full boost to 4950+ on some cores. I was screwing with the power profiles and it didn't matter, maybe its my OS. Idle at 55-60C and 50W is a bit high for stock no load operation. I didn't spend that much time on it yet, but it appears stable at 4200/4150 at 1.075. Thermals are fine like this, but no boost is kinda silly, and there's still no thermal limit. The computer won't turn off or throttle when it exceeds 105C before I stopped P95 manually. Isn't it supposed to be 95C? It would be nice if I can just manually define steps above base clock and it would obey.

  6. #6
    Xtreme Mentor
    Join Date
    May 2008
    Location
    cleveland ohio
    Posts
    2,864
    I don't know something doesn't seem right.
    It's not throttling at all above 95C ?
    I would see if I could RMA it, seems like something inside of it doesn't work right.
    HAVE NO FEAR!
    "AMD fallen angel"
    Quote Originally Posted by Gamekiller View Post
    You didn't get the memo? 1 hour 'Fugger time' is equal to 12 hours of regular time.

  7. #7
    Brilliant Idiot
    Join Date
    Jan 2005
    Location
    Hell on Earth
    Posts
    11,112
    Quote Originally Posted by ZOMGVTEK View Post
    Dang. Thanks for the info. I had it running a manual OC since something odd is going on and it never goes below full boost at idle. I was getting around 1.5V claimed Vcore at idle, 0% load, full boost to 4950+ on some cores. I was screwing with the power profiles and it didn't matter, maybe its my OS. Idle at 55-60C and 50W is a bit high for stock no load operation. I didn't spend that much time on it yet, but it appears stable at 4200/4150 at 1.075. Thermals are fine like this, but no boost is kinda silly, and there's still no thermal limit. The computer won't turn off or throttle when it exceeds 105C before I stopped P95 manually. Isn't it supposed to be 95C? It would be nice if I can just manually define steps above base clock and it would obey.
    yep crosshair literally shuts down @ 95c no throttling with a fixed OC. 50c is ehh maybe a touch on the high side although i think i was around 45c on my 5900x on aio 280 mm. they idle fairly warm. stock operation i observed chip throttles back and holds temps at 95c in prime.
    heatware chew*
    I've got no strings to hold me down.
    To make me fret, or make me frown.
    I had strings but now I'm free.
    There are no strings on me

  8. #8
    Xtreme Enthusiast
    Join Date
    Feb 2008
    Posts
    832
    I messaged ASUS a few days ago to see if they can figure it out. No word back yet.

  9. #9
    Xtreme Enthusiast
    Join Date
    Feb 2008
    Posts
    832
    I got a response, but they basically said they remove all limits when overclocking to eliminate erratic behavior, and if you want to change the temp limit you can do so. Trouble is my board entirely ignores the temp limit setting they claimed should work, and the current limits are not active as well. So I suppose it is what it is.

  10. #10
    Brilliant Idiot
    Join Date
    Jan 2005
    Location
    Hell on Earth
    Posts
    11,112
    I have yet to get a succesfull pass with 5950X in prime that does not end up me waking up with a cpu over temp bios screen. I've seen it 4-5 hours in fine and wake up in morning to that. using a liquid freezer II 280 mm now and just PBO. I may set the EDC lower now as It's 100% of the 200a always......maybe that will fix it.

    All the other sku cpu's are tamed now for using curve optimizer and PBO with the liquid freezer 280. Unsure about manual but i get better results with curve optimizer and PBO anyway.

    I guess anyone serious enough to run a 5950x should probably run a triple rad or custom loop to get the most out of it.
    Last edited by chew*; 12-28-2020 at 10:41 AM.
    heatware chew*
    I've got no strings to hold me down.
    To make me fret, or make me frown.
    I had strings but now I'm free.
    There are no strings on me

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •