Asrock X370 Bios Update
Kristin Klodzinski
I just finished a brand new Ryzen build. It runs very well under load, but freezes randomly at idle. By "freezes" i mean it's like everything stops responding instantly, mouse doesn't move, no BSOD, no restart, no graphical artifacts.
System temps are very well within normal parameters. Cooling is provided by 3x140mm Noctua fans connected directly into the MB (So i can monitor their RPM and adjust PWM curve), CPU is cooled by an NH-U12S SE-AM4. System voltages from the PSU are rock-stable too - the Seasonic doesn't break a sweat.
This is a system which I intend to use for my daily work and stability is of utmost importance for me. I haven't overclocked anything - everything is at stock "Auto". I even down-stepped the RAM from the XMP-3200 profile to the safe JEDEC 2400Mhz.
The funny thing is it rock-stable on all the stress tests i threw at it. 3DMark, CPU-Z, Aida64, Heaven Benchmark, and several hours of Memtest86. It's just this totally random hard lock screen freeze (or whatever you can name it) when I'm basically doing nothing - like writing in a text document or going around in the Control Panel, or even when i'm not at the computer at all. There's nothing in the Event Log (besides the improper shutdown message), no crash dumps either.
Done a full CMOS clear and haven't touched anything but the Fans PWM curves and the Dr.Debug led. I'm expecting the MB to honor it's "Auto" values and not apply any additional overclocking, besides the Core Boos and the XFR, which are some of the highlights of this CPU.
[My company is using only AMD as we feel the need to fight Intel's monopolistic behavior, we have more than 200 AMD based computers most of them self-built, all of them running Linux (as we feel the need to fight Microsoft's mono.......).
Initial testing with our own workload/software and standards like Blender showed nothing noteworthy, everything was working (fast) as expected. As outlined in the original post, however, we had "random" freezes - usually when the machine was unaccompanied and probably mostly when it was idle. At first this actually went overlooked and was attributed to remote power failures or the like.
Recently the freezes became more frequent even though nothing in the use of the machine changed. Very recently the machine started freezing "even" when used actively - to the point that it became basically useless.
I started looking into it - temperatures, RAM and all that - but couldn't find anything. I contacted AMD through the support line and after a few attempts to jointly troubleshoot (disable C6 and some others) RMA'ed the processor. Since I need the machine I couldn't wait for the proc to be returned and bought another one, in this case a Ryzen 3, 1200 (I know, very different - this will be used to repurpose some parts).
The freeze looks the same as the ones before: Machine just locks, graphics still up, monitor on - but nothing moves. No remote login (ssh) possible. It looks like it happened again when the machine was idle.
PC Freezes randomly, no matter what I do. With different cycles, from 5 offsets per day, up to 1 shutdown per week. Disables mostly when I work in Blender 3D, but sometimes when I just listen to music or use the internet. When conducting stress tests did not turn off. All software is licensed, including Windows 10 Home, viruses. Problems began in early 2018.
Thank you. I've tried to do this before. Did not help. Also made a post PC Freezes randomly. Ryzen 1700. B350. RX 460 Posted to technical support. They contacted me, asked for more information, I sent them all the necessary information, waiting for an answer. I think that in our case there are two problems that give the same symptoms. The first is a problem with C6, which appeared to me in January 2018 after updating BIOS. And the second is a problem with the GPU, which apparently decided, publishing new drivers a few weeks ago.
I can't give you any advice if your system was freezing under load, but if the problem was with the C6 states, try disabling the "Core Parking" feature in Windows. This should prevent the CPU to enter C6 state:
If I understand correctly from my own tests, disabling the "Global C-States Control" would also disable Coo'n'Quiet, which, for a stock system, is not that desirable - you wish for the CPU to lower its voltages and frequency when not loaded. I might be wrong though, and I stand corrected.
Deducting that after the black screen you still hear music playing for a few seconds, I can conduct that the CPU is still working at that point. This indicates to me that there is a problem either with your GPU (or drivers), or there is a problem in the BIOS. Again, this is only an assumption and I stand corrected again.
I turned off link state power management(it is like fix the problem but I dont use this for long peroid of time) but it lock the cpu clock to 4.075 or 4.050 Gigahertz on all cores that I see ryzen master ans then I turned off?
My pc working good after I disconnect my keyboard until I plug my phone and it shows error in windows and device manager it not recognized my phone(I fix the not recognized problem after reinstalling samsung driver again I think it shows error because I need to restart for driver properly however I dont restart pc for reinstall but this time it detect without error my phone when in downloading) and I dont know it is unplug pluging without reject problem or my motherboard sensitive for problem devices
just I want to know can mouse and output device not working permanently(all output devices like keyboard and mouse not work and screen stable until restart from device problem that windows 10 or motherboard)?
AMD sent me a replacement Ryzen 7 1800x but the problem of random freezes persists. Interestingly, it is less frequent than with the original processor. For a while I was using a Ryzen 3 in that machine (ARock x370 gaming board) and that also froze a few times.
A second system based on an AsRock ITX (x370) and a Ryzen 5 2400G is quite stable but froze once so far - while watching a Youtube video (maybe it was the content, something about Trump - just kidding...)
I tried multiple memory combinations - it is possible that the frequency of freezes correlates with the amount of memory in the machine. I have distributed my (expensive) memory to the two machines instead of having it all in one machine and the freezes have become less frequent - might be related or not, can't say.
After reading various forums, I saw that those who overclock the system have a stable system (most of them overclocking). Those users who are all standard have problems. I assume that the problem is in "AGESA", which is in every bios. It does not have the right standard CPU power settings. The power goes down to a lower critical level and the links between the CPU and the RAM or the PCI lines are lost. Two days ago, I turned off the "global C states control", and yesterday got a blue screen that pointed to the RAM problem (checked by memtest86, there are no problems). And before that, at first the image disappeared, during this GPU the cooler worked. That is, the link between the CPU and the GPU is lost, and power is supplied to the GPU. We have different chipsets, motherboard manufacturers, processor models, and the problems are very similar. Sorry for my english, I hope you all understand what I wrote
The last thing I did: installed the driver video (amd) from windows update center.In Bios set the default settings. The problems have disappeared. Then windows was updated and installed the latest version of the bios. If it does not help, try installing another power supply.
Somebody writes about hardware issues (I also thought that before). And now I think that this is a software problem. Because when I bought a new pc, everything worked perfectly. The problems occurred after reinstalling windows to a newer version. At a time when windows had permanent crashes and did not help anything. I installed a Linux because I needed to work, and I had no problems with it. Now I use windows again as the main system, 17 days everything works perfectly
I had the exact same issue! Everything worked perfectly until I installed the latest windows 10 update and the newest BIOS update on the same day. A couple of days later, the crashing started at idle. Now i don't know what is causing it. I even downgraded to the old BIOS but the crashing problem at idle has stayed. I even re-installed windows and sometimes it crashes during windows install. It's so random(mostly during idle) and i can't reproduce it always.
I have had the AMD RMA'ed 1800x up and running for 12 days now. Tested some parallel blender runs in the first few days but then idle for most of the time. Since it's been up all this time, obviously no freezes/crashes at all.
I had to set an option in my BIOS called "Power Supply Idle Control" to "Typical Current Idle". I've never had a single freeze since then.
I have an ASRock motherboard, so i'm not sure if such an option exists in Gigabyte or how it's called.
Check my thread at ASrock forums for more info:
1800x & x370 Taichi Freezing issue - ASRock Forums - Page 1
The Guaranteed way ( for me) to cause the freeze/fail is by compiling LLVM source code 8.0 &/or 9.0 with ninja as the compiler on all platforms. Im assuming( in my case scenario ) that this is due to the default amount of threads utilized.
c80f0f1006