What is going on? Im reading WAY too many people with idle freeze problems with 1st gen Ryzens

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

kschendel

Senior member
Aug 1, 2018
261
190
116
...
The good news is that I’ve read that the latest updates for many boards, which is BIOS 5.50 for the X370 Taichi, finally solve the freezes. Unfortunately, this BIOS for the Taichi also introduces a nasty new bug : if my understanding is correct, if you select CPU for the fan control source, it now monitors the Tdie, which varies a lot, instead of the CPU sensor on the mainboard. I think I’m going to wait before update the BIOS

That's not a bug, it's a feature. The CPU sensor on the board itself is broken (stuck) on a wide variety of Asrock boards. I don't have a handle on whether it's some boards or all boards, but I certainly have one (X470 Taichi ultimate where the board CPU sensor reads 40C no matter what). I've no idea what the underlying sensor issue is, but apparently it's not fixable from the BIOS, so they just disabled it.
 

Val_

Member
Nov 24, 2012
29
0
66
That's not a bug, it's a feature. The CPU sensor on the board itself is broken (stuck) on a wide variety of Asrock boards. I don't have a handle on whether it's some boards or all boards, but I certainly have one (X470 Taichi ultimate where the board CPU sensor reads 40C no matter what). I've no idea what the underlying sensor issue is, but apparently it's not fixable from the BIOS, so they just disabled it.

I’m very surprised at this ! Did someone from ASRock tell you this ? The CPU sensor on my board clearly isn’t stuck. Removing this option is terrible :/
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,839
3,174
126
I dont think this is a AMD only thing.

I am getting the so called "IDLE FREEZE" as well from time to time.

I have adjusted my overclock to stock settings and its reproducable.
Im 100% sure my system is rock stable. I have ran Burntest for several hours, and even memtest.
Its not stability on my system, but its most definitely software related.
It doesn't always freeze, but something causes it to freeze and i cant seem to figure out what it is besides the recient windows/nvidia update.

My system is not an AMD but an INTEL i9-7920X

So at this point im blaming microsoft or NVidia for FUBARing something, as i started getting these after i have updated both on my system.
 
Last edited:
  • Like
Reactions: Markfw

ZGR

Platinum Member
Oct 26, 2012
2,052
656
136
I had a terrible idle freeze issue on my GTX 1070 late summer 2016. This was after I upgraded from an R9 290 using DDU. I ended up reinstalling Windows.

If I was still suffering from this, I probably would have tested another OS. Windows is definitely not bug free.
 

kschendel

Senior member
Aug 1, 2018
261
190
116
I’m very surprised at this ! Did someone from ASRock tell you this ? The CPU sensor on my board clearly isn’t stuck. Removing this option is terrible :/

There were so many reports of stuck CPU sensors on the Asrock forums that it eventually got turned into a sticky thread. I think we can assume that that was a tiny fraction of the RMA's and support issues. And as I said, mine was clearly broken.

It's interesting to hear that at least one person had a working CPU thermistor sensor, makes the whole situation even more curious.
 

Hi-Fi Man

Senior member
Oct 19, 2013
601
120
106
This issue has been PLAGUING me for days now. I'm running an ASRock B450 ITX board and Ryzen 3700X. Latest firmware (3.50) too. No matter what I do my system will randomly freeze on idle. I tried setting common current idle to enabled, but no dice. I'm testing if using the Ryzen high performance profile fixes this and if not I will try disabling global C-state control (what exactly does that do?). It seems like a recent Windows update F**ked something up.
 

HisEvilness

Member
Mar 23, 2019
34
2
16
www.hisevilness.com
I have a 1600x and i had a 1600 the problem is with the DDR4 it is a bit finicky with 1st gen Ryzen, you want B-Die(Samsung) or A-Die(Hynix) single rank sticks, I had dual rank and it just randomly BSOD's even when running 2133/2400 while the sticks are rated for 3200. When it hangs it means it is almost stable, try a slight volt increase or set the memory to run one step slower under XMP.
 

misuspita

Senior member
Jul 15, 2006
388
417
136
This issue has been PLAGUING me for days now. I'm running an ASRock B450 ITX board and Ryzen 3700X. Latest firmware (3.50) too. No matter what I do my system will randomly freeze on idle. I tried setting common current idle to enabled, but no dice. I'm testing if using the Ryzen high performance profile fixes this and if not I will try disabling global C-state control (what exactly does that do?). It seems like a recent Windows update F**ked something up.
I had problems with my AsRock AB350M Pro4 freezing randomly with my 1600. Tried everything, entered 2 times in repairs (warranty). Second time the gave me my money back, it was something with the board. Replaced with a AsRock B450M Pro4, works like a champ, no problems. Oh, and my RAM couldn't reach higher than 2666 in AB350m, in B450 got direclty to specified freq, meaning 3000
 

Hi-Fi Man

Senior member
Oct 19, 2013
601
120
106
Crashed again! Disabled global C-state control and we'll see what happens. One thing I noticed, and this just started very recently, but HWinfo64 crashes very consistently now with just being open. So far with global C-state control disabled HWinfo64 does not crash...
 

Hi-Fi Man

Senior member
Oct 19, 2013
601
120
106
No lockups so far. Looks like C-states are bugged right now (at least for this board). Easiest way to verify whether it's stable or not is to just open HWinfo64. With global C-states enabled, HWinfo64 will crash/close within a minute. This has got to be the buggiest platform I've ever dealt with, and that's going back as far as socket A and socket 478 days.
 

KaiserARM

Junior Member
Sep 11, 2020
2
0
6
Help meeee. I have the same problem. Freeze on idle. Monitor went blank. Mkb wouldn't respond. Power at full performance.

Im running 3990x. 256GB ddr4 3600mhz downclocked to 3533mhz (otherwise this pos wouldnt post pun intended), 1080 ti sli.

So this is not just first gen tr problem.
 

Gideon

Golden Member
Nov 27, 2007
1,608
3,573
136
Help meeee. I have the same problem. Freeze on idle. Monitor went blank. Mkb wouldn't respond. Power at full performance.

Im running 3990x. 256GB ddr4 3600mhz downclocked to 3533mhz (otherwise this pos wouldnt post pun intended), 1080 ti sli.

So this is not just first gen tr problem.
Could you try running the memory @ stock, just to debug if it still happens?

Are you running custom timings for the memory or XMP? If only XMP it might be that secondary timings don't actually play well with your CPU/Memory combo. My nephew's 3800X had a similar issue (his XMP profile @ 3600 Mhz wouldn't even boot) and setting custom timings solved the issue and offered a nice perfomance boost on top.

This guide is a decent starting point:

The tool for calculating correct timings is this (also works for threadripper, but make sure you select the correct options):

You'll need a memory test to validate that your memory is stable (you can run that at 3533 as well to validate it's actually stable). If you're running windows (which mighy not be the case considering the CPU) I recommend Karhu MEM-test for that. It costs, but it finds the errors super quick (usually passing 100% means you're good), while other memtests need hours to run. If you're on Linux then this tool should also work.

Also obviously always try the latest BIOS and latest OS version (kernel).
 
  • Like
Reactions: Shmee and Elfear

kschendel

Senior member
Aug 1, 2018
261
190
116
Try setting power supply idle current to Typical current idle in the BIOS. The other bit of BIOS rubber-chicken-waving you can try is disable C6 sleep, or if you can't find that, disable global C states.
 
  • Like
Reactions: therealmongo

Shmee

Memory & Storage, Graphics Cards Mod Elite Member
Super Moderator
Sep 13, 2008
7,381
2,415
146
Yeah, I would definitely memtest that. I like the UEFI passmark memtest86, free version. Runs for up to 4 passes. Do that and check for errors. Note that this is a bootable test.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,478
14,434
136
I can't remember where I saw it, but I don't think that motherboard supports that speed with 32 gig dimms in all 8 slots. I think it was 2933 or 3200 max
 

richaron

Golden Member
Mar 27, 2012
1,357
329
136
Help meeee. I have the same problem. Freeze on idle. Monitor went blank. Mkb wouldn't respond. Power at full performance.

Im running 3990x. 256GB ddr4 3600mhz downclocked to 3533mhz (otherwise this pos wouldnt post pun intended), 1080 ti sli.

So this is not just first gen tr problem.
Official site says "System Memory Specification Up to 3200MHz"
edit: woops copypasta & numerous edits but you get the point I hope. Another case of jumping on the bandwagon to blame someone else. Nobody likes to admit it's their own fault.
 
Last edited:
  • Like
Reactions: Markfw

KaiserARM

Junior Member
Sep 11, 2020
2
0
6
Could you try running the memory @ stock, just to debug if it still happens?

Are you running custom timings for the memory or XMP? If only XMP it might be that secondary timings don't actually play well with your CPU/Memory combo. My nephew's 3800X had a similar issue (his XMP profile @ 3600 Mhz wouldn't even boot) and setting custom timings solved the issue and offered a nice perfomance boost on top.

This guide is a decent starting point:

The tool for calculating correct timings is this (also works for threadripper, but make sure you select the correct options):

You'll need a memory test to validate that your memory is stable (you can run that at 3533 as well to validate it's actually stable). If you're running windows (which mighy not be the case considering the CPU) I recommend Karhu MEM-test for that. It costs, but it finds the errors super quick (usually passing 100% means you're good), while other memtests need hours to run. If you're on Linux then this tool should also work.

Also obviously always try the latest BIOS and latest OS version (kernel).


this just started recently. ive been running 3533mhz for ever. recently I added third hdd and this started to happen. yes ill try everything.

people on amd group in facebook even running their ram to over 3600 with no issue. (not threadripper tho)
 

teejee

Senior member
Jul 4, 2013
361
199
116
this just started recently. ive been running 3533mhz for ever. recently I added third hdd and this started to happen. yes ill try everything.

people on amd group in facebook even running their ram to over 3600 with no issue. (not threadripper tho)

3533 was probably very close to the limit before, and for some reason it is now just too high. This is not strange, there are many reasons the capability can change a little bit (e g the temperature changes causes mechanical changes in the PCB/connectors).

Change mem speed to around 3400 and be happy with your build :)
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,478
14,434
136
this just started recently. ive been running 3533mhz for ever. recently I added third hdd and this started to happen. yes ill try everything.

people on amd group in facebook even running their ram to over 3600 with no issue. (not threadripper tho)
You may have run it at that, but why are you upset that it won't continue to run at that speed, when it is clearly over the AMD platform approved spec'ed and design speed ? Its clearly stated that its 3200 on the AMD 3990x website. Here in this post https://forums.anandtech.com/thread...ems-with-1st-gen-ryzens.2562813/post-40280454