PSA: testing for and solving random freezes on Ryzen

Hardware CPU
22

In that case I will also link this thread again that I put into the PM. The rest of my thoughts I have expressed there. :slightly_smiling_face:

23

Wasn’t that a problem with Haswell (?) at one point? Some power supplies couldn’t handle the low power draw and their voltage protection would trigger, causing them to reset.

EDIT: I shouldn’t say same, but a similar-ish problem.

1 Like
24

Not iust haswell a lot of OEM systems. Out of the box premiere crashing under load is a thing. It was fixed via updates and microcode which boost the voltage slightly. It’s an ongoing problem. On the dells the temporary fix was also to disable power management but it’s a different issue than ryzen. Disabling power management is always a thing.

I would say maybe something like if after disabling all power management the sytem still intermittently locks up you should look at your Mobo ? Hah?

1 Like
25

I’d certainly agree with that.

But to just generalize and say “X problem is ALWAYS Y” (as per @SamX above) without any actual verification/validation is how you end up with broken diagnosis (and possibly say, RMA a good board and wait for replacement when the problem is actually something entirely different - wasting both your time, the vendor’s time, etc.).

Develop a hypothesis, test, confirm result.

3 Likes
26

I have a laptop with Raven Ridge (2200U). It locks up in Linux after a set time if I don’t turn off C6-states with zenstates.py. Windows can run longer without locking up.

3 Likes
27

Yeah kinda i guess.
Atleast start to disable all power saving features and go from there.

1 Like
28

Hello, sorry for bumping this thread up again, i am kinda lost and hope you guys can help me out here.

I experience freezes on a 2700X X370 Taichi system.

The system ran fine for weeks with 4*8GB G.Skill SniperX Digital Camouflage at 3200MHz. Then i went and installed one 8GB Kingston KSM24ES8/8ME ECC and everything was fine as well. After adding three more it went bad. I tried all three of the new ones with the one that ran fine for weeks and all froze (2 sticks total). I tested the RAM on said system and another X399 system for 2,5h each without any errors. I am on the newest recommended BIOS version 5.60 for that CPU and also tried bumping the DRAM voltage from 1.2 to 1.22. I could not find the power supply idle control option in my BIOS nor did i find anything saying C6 but i set 3 manual frequency states being 3600MHz, 3200MHz and 2200MHz. Is it sufficient for it not to draw too little power?

Can you point me to any other settings i could try? Is the RAM bad although it ran fine in memtest86?

In summary:

  • 32GB G.Skill SniperX @ 3200MHz (4 sticks) is fine
  • more than 8GB of Kingston ECC (1 stick) @ 2133/2400 freezes

Thanks in advance, cheers

29

Do you have the gear down mode option ? If so disable that.

You could run that kit as high as 1.35 without issue.

30

Thank you for the prompt reply! I will look into that tonight. Would you recommend running 24/7 at this voltage?

31

I might have read somewhere that later UEFI versions removed the Power Supply Idle Control option. If you really cannot find it where I described it should be, than you may have to downgrade your firmware. Unfortunately downgrading on this board is not officially supported and may be a huge pain in the ass to do without an external flash programmer.

32

@wendell

So i set the voltage to 1.335 which shows as 1.352 in the hardware monitor and also disabled gear down mode. Lets see how it plays, 3 days are to beat.

@methylzero

Thanks for your input, i hope though i don’t have to go doen that road.

33

That should be ok 24/7 for at least a few years.

34

I am fine with that, should it be stable now i’ll try to find the minimum stable voltage. Maybe even just the gear down mode was the culprit.

So far it looks good, day number 2.

On another X399 system i could find the settings advertised in the first post and i also disabled gear down mode there as well.

I am not entirely sure but i think since that this machine does not occassionally stall when running a mac VM anymore.

35

@wendell unfortunately it froze again, do you have some other settings to try?

36

You said you didn’t see an option for c6/7 stare? What about global c states? If you have that set it to disabled?

Also:
Advanced > AMD CBS > Zen Common Options > Power Supply Idle Control
This should be set to Typical Current Idle .

Do you not have this option??

37

No didn’t see that, global c states are disabled. I now put back in the old RAM which was working fine and now even that one freezes. It must be the board then.

38

What video card do you have? Ever since I upgraded my 1060 to a 5700xt I have been getting freeze every once in a while. I narrowed it down to the card, because once I switched back to my 1060 it quit happening.

39

It is a Radeon HD R5 230 (device-id: 0x6779). I am not sure if updating to 5.60 was a good idea. It was perfectly fine before and now i cannot get the original setup working without a freeze. That same graphics card was in there before. I guess i’ll try to reset the BIOS and if nothing helps get a new motherboard.

40

Also is the dude pulling power from wall or extension cable shared with other devices?

41

If you have a good PSU the extension cable should not matter. BTW a flaky PSU may also cause random freezes.

You can try using ZenStates to disable Package C6. I think it is more or less equivalent to the BIOS setting you cannot find. If that does not help you might have to jump though the hoops and downgrade to 4.xx BIOS version where you still have the BIOS option that helped me.