cancel
Showing results for 
Search instead for 
Did you mean: 

When we we see a decent bios update so that ddr5 runs in XMP ?

Undertoker
Level 9
All I’ve experienced with a hero and an extreme z690 board thus far is annoyance and aggro
When will Asus release a bios that actually confirms to its advertised qvl listing ?*
You guys really need to get your act together frankly , it’s damaging your brand *
I9 14900K >ASUS ROG Strix RTX 4090 OC Edition
ASUS ROG MAXIMUS Z690 EXTREME EATX Motherboard
ASUS ROG PG35VQ @ 200mhz in 3440x1440
Trident Z RGB 64GB DDR5 @ 6400mhz
Samsung 980 EVO Pro M.2 x 5 off
NZXT Z63 Kraken AIO CPU Cooler
NZXT H700i + Hue 2 (Pubg Edition) - Heavily modified
ASUS ROG THOR 1200w PSU
Corsair M65 RGB Mouse
Corsair K95 RGB Keyboard
8,252 Views
45 REPLIES 45

trihy wrote:
Freezes at idle are rarely related to ram.

There is a problem with Asus bios applying too low voltage to cpu at idle.

Just use hwinfo and you will see cpu vcore voltage dropping to 0.300v from time to time. This is way out of specs and gives random freeze.

Using voltage offset, doesn't matter if + or -, will give more frequent random freeze.


That might explain why disabling c-states seems to bring more stability to a system. I wonder if there's an alternative approach for setting just the minimum voltage, so some power savings could still be realized. Possibly a combination of Package C-State Limit and Enhanced C-States?

Update: monitoring CPU voltage with HWinfo64, I've seen it drop as low as 0.204v!

0.222v is... Wow that's really low. No wonder there's freezing. Admittedly I didn't think to monitor CPU vcore voltage, this could be potentially enlightening. Yes, ideally having some power savings with C-states enabled would be desired, though I think I don't mind doing without till we have a less absolute solution. I've got C-states disabled but I think it might be worth it now, just so we can know for certain, re-enabling and logging HWInfo for a day or two to see how low it gets.

- Edit: With C-states disabled, HWinfo is reporting a 1.279v - 1.288v. Very little fluctuation, definitely within the 1.25v - 1.280v specification, maybe a bit over, for the 12900k. I'll try re-enabling C-states well, this needs investigation. Yeah I hope Asus takes note of this and addresses it with the next version.

trihy
Level 9
Most user are blaming ram, but rams gives more bsod than freezes.

Random freezes at idle are mostly caused by cpu. And those crazy low voltages are not normal. Hope Asus take a look into this.

Not sure what setting could rise idle voltage. It's not near at all from Intel specs. Tried fixed cpu core voltage and kind of works, but far from ideal.

trihy
Level 9
Remember to check cpu core voltage. Not cores voltages. On hwinfo. Only cpu core voltage (vcore) is the real voltage the cpu is getting. The other ones are the ones the cpu is expecting per Intel voltage table.

trihy wrote:
Remember to check cpu core voltage. Not cores voltages. On hwinfo. Only cpu core voltage (vcore) is the real voltage the cpu is getting. The other ones are the ones the cpu is expecting per Intel voltage table.


Yup, noted, with thanks. I'll start logging tomorrow.

trihy wrote:
Remember to check cpu core voltage. Not cores voltages. On hwinfo. Only cpu core voltage (vcore) is the real voltage the cpu is getting. The other ones are the ones the cpu is expecting per Intel voltage table.


Yep, Vcore is the one I've been graphing all afternoon. It averages 0.748v, but the lowest point it's reached is now 0.204v. The highest it's been while I've been graphing is 1.376v, probably from a prime95 or Cinebench run.

92926

trihy
Level 9
Yes, I'd noticed my cpu seems to tolerate 0.3xx most of the times, but when things gets around 0.2xx it will sooner or later freeze. Most of the time when I leave the PC at idle, come back after a few hours and it's freezed.

Sometimes it happens more than one time per day. Other times it can get around a couple of days before freezing. I guess better binned cpu can tolerate best this crazy low voltage.

trihy wrote:
Yes, I'd noticed my cpu seems to tolerate 0.3xx most of the times, but when things gets around 0.2xx it will sooner or later freeze. Most of the time when I leave the PC at idle, come back after a few hours and it's freezed.

Sometimes it happens more than one time per day. Other times it can get around a couple of days before freezing. I guess better binned cpu can tolerate best this crazy low voltage.


I just watched mine briefly hit 0.009v after a restart with the machine completely idle, but shockingly, the system did not freeze. This makes me kind of question the accuracy of the readings in HWinfo64 though, or our assumptions about what a sane range of voltages should be for the 12900K. If ultra-low voltages should be expected to trigger a hang, surely 0.009v should have done it.

I think I'll leave the hwinfo64 CSV logger running; there might be something interesting to see if/when my system hangs again.

trihy
Level 9
That could be a read error.

But if you keep watching vcore, you will notice it goes frequently to 0.3xx or 0.2xx values. That doesn't seem like a read error. Tried some fixed voltages and hwinfo always show the correct one.

That could explain also why setting -0.050 offset makes the system hangs at the Asus logo. All alder lake cpu support more than that. So probably the hang at Asus logo it's because the crazy low voltage plus the voltage offset, a killer combo for cpu idle voltage.

Also tried OCCT and show the same 0.2xx and 0.3xx problem.

If you have a friend with alder lake and other brand motherboard, ask to check vcore. This could answer our clues 😄

trihy wrote:

If you have a friend with alder lake and other brand motherboard, ask to check vcore. This could answer our clues 😄


I'm certainly going to be tempted to buy another brand motherboard if this thing freezes up randomly one more time. I'll let everyone know what I learn if/when I do.

(I'm still trying to rule out external hardware and third-party software on mine.)