Everything freezing on screen then PC restarts

[Healadin]

Hello,

 

I already tried discussing it here and the solution even worked for a while, but to describe my problem it is a bit of a story time. Firstly I will start with my config:

 

-CPU: Ryzen 7 7700X

-GPU: ASUS DUAL Geforce RTX 4070 OC

-MB: ASUS ROG STRIX B650E-E GAMING WIFI

-Cooler: ASUS ROG STRIX LC II 360 ARGB

-SSD: Samsung 980 Pro 2TB

-PSU: ASUS ROG STRIX 1000W Gold Aura Edition

-NIC: ASUS XG-C100C V2

-fans: Lian Li UNI FAN AL120 V2 ARGB PWM x7

-OS: Win 10 Pro

-RAM: 2x 16 GB Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32

or 4x 16 GB Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32

or 2x 16 GB G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30

or 4x 16 GB G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30

 

And I mentioned RAM as last since it is suspected culprit, at least in the last thread that I had here. I started with 32 GB (2 sticks) of Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32. It didn't look best in the case (Lian Li O11 Dynamic XL, ROG certified), so I decided to buy another 2 sticks. But computer started acting weird because of 4 sticks at 6000 MHz, people here advised me to not use DOCP profile with 4 sticks, since officially supported speeds for 2 sticks are only 3600 MHz, slowly I started overclocking the RAM manually and found out that 5600 MHz 30-36-36-75 worked and was stable (Prime95 for 10 hours, Memtest86 for 4 hours). And it was true, everything ran nicely for at least 2 weeks. Then out of nowhere Windows update came up, which I didn't even applied, it just applied itself (pressed shutdown instead of "shutdown and update" option and update started anyway). And since then everything is screwed again.

 

I play Runescape on one screen (or sometimes both with alt account) and watch youtube in Firefox windowed mode or just listen to it on background and out of nowhere everything freezes, I can move cursor, but basically clicking does nothing, audio from youtube runs for like another minute or so, then audio stops and within minute windows restarts itself. In windows event logger there is just "The system watchdog timer was triggered." and "unexpected reboot", that's it.

 

And now for everything I tried until now. Using default profile with all 4 Kingston sticks. That did not do the trick. Then I decided to switch to G.Skill memory, since these kingston kits are just "XMP certified", but not EXPO certified. I also bought 2 kits of these G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30 16 GB sticks (2 in kit, 4 overall). I read in reviews, that people had no problem running 4 of them at 6000 MHz, so I just put them in and activated DOCP. With benchmarks (Prime95 and  Memtest86), it ran without errors, but in real use it didn't work for me - random freezes again. So next thing I tried was to run just 2 sticks of Kingston memory at 6000 MHz CL32. Started with benchmarks (another almost day), again everything ran without errors, so I started using PC again. It took maybe 2-3 hours, I went for dinner, returned to PC and it was rebooted. So another failed attempt. Then I tried the last option that definitely should have worked and switched to 2 sticks of G.skill EXPO certified memory at 6000 MHz CL30. It took roughly 6 hours of use to get first random freeze.

 

So at this point I am running with 2x 16 GB G.Skill Trident Z5 neo RGB DDR5 6000 MHz CL30, which should be AMD EXPO certified and with almost every configuration (except 4x Kingston at 6000 MHz CL32) all benchmarks run without errors - Prime95 (Large FFT - stresses memory controller and RAM) for 8+ hours, Memtest86 for 3-4 hours (max number of runs in free version). With all the swaps I stopped installing Fury CTRL software/Trident Z lighting controll and just use aura software from armoury crate and I have installed L-Connect 3 for the LianLi fans.

 

I am also running latest Bios (1616) and all chipset and drivers are up to date (using Armoury Crate and Geforce Experience to update them).

 

I was thinking maybe upgrade to windows 11 might do the trick, but honestly I prefer Win10, so haven't done it yet. If someone has any suggestion what else I can try, feel free to suggest it please, honestly I am starting to be kind of hopeless at this point.

[Healadin]

so no one has any idea, what might be happening?

[Damocles Ti]

Platform is new, not a lot of people have what you have to give advice. Memory is a very likely issue with the new platform, and the new socket doesn't help either. If you're experiencing random reboots. I'd start with checking event viewer for system logs; we're going to need a lot more logging and context to narrow down the issue. I'd uninstall armory crate as that thing meets the definition of a virus. Random reboots are either RAM or CPU. Is the CPU overclocked, PBO?

 

My best guess for freezing is memory, but rebooting is usually CPU or power limits on motherboard

[Healadin]

1 hour ago, Damocles Ti said:

Platform is new, not a lot of people have what you have to give advice. Memory is a very likely issue with the new platform, and the new socket doesn't help either. If you're experiencing random reboots. I'd start with checking event viewer for system logs; we're going to need a lot more logging and context to narrow down the issue. I'd uninstall armory crate as that thing meets the definition of a virus. Random reboots are either RAM or CPU. Is the CPU overclocked, PBO?

 

My best guess for freezing is memory, but rebooting is usually CPU or power limits on motherboard

I have only activated DOCP (Ram overclock), PBO I have not activated, but it is 7700X, so its the 105W version of chip, so probably its active out of the box?

 

anyway here are the critical logs, havent found export option quickly, so just made some screenshots

 

[Damocles Ti]

3 minutes ago, Healadin said:

I have only activated DOCP (Ram overclock), PBO I have not activated, but it is 7700X, so its the 105W version of chip, so probably its active out of the box?

 

anyway here are the critical logs, havent found export option quickly, so just made some screenshots

 

Unfortunately these errors show nothing, and usually a CPU issue would put something in event logs half the time. I would also suspect it's RAM or memory controller. You can install something like HWinfo64, there's a WHEA listing there that loves to count up every time the memory controller is struggling to keep the FClock stable for your ram kit. Prime95 is not very good at testing memory but is excellent for testing Curve Optimizer stability (which we're not doing), and Memtest isn't good for testing something that's edge of stability either. I've found that Aida64 extreme's memory stress test works best at finding these barely stable memory overclocks.
 

• Try running HWinfo64 on a second monitor and check if the last row "WHEA" increases with your computer under load or gaming. If it does, your CPU's memory controller cant handle the RAM speeds. Might need SOC voltage tuning but that's a different rabbit hole
• Install Aida64 (trial version) and run the memory stress test under tools. It doesn't take very long, this baby usually finds errors within 10 minutes but I run it for an hour to be sure. If it quits, it means the memory OC isn't stable enough for regular use.

[Healadin]

12 minutes ago, Damocles Ti said:

Unfortunately these errors show nothing, and usually a CPU issue would put something in event logs half the time. I would also suspect it's RAM or memory controller. You can install something like HWinfo64, there's a WHEA listing there that loves to count up every time the memory controller is struggling to keep the FClock stable for your ram kit. Prime95 is not very good at testing memory but is excellent for testing Curve Optimizer stability (which we're not doing), and Memtest isn't good for testing something that's edge of stability either. I've found that Aida64 extreme's memory stress test works best at finding these barely stable memory overclocks.
 

• Try running HWinfo64 on a second monitor and check if the last row "WHEA" increases with your computer under load or gaming. If it does, your CPU's memory controller cant handle the RAM speeds. Might need SOC voltage tuning but that's a different rabbit hole
• Install Aida64 (trial version) and run the memory stress test under tools. It doesn't take very long, this baby usually finds errors within 10 minutes but I run it for an hour to be sure. If it quits, it means the memory OC isn't stable enough for regular use.

Can you navigate me a bit please? I cannot find this WHEA counter

[Healadin]

9 minutes ago, Healadin said:

Can you navigate me a bit please? I cannot find this WHEA counter

ah found it, Windows Hardware errors (WHEA)at the very end of sensor window

 

[Healadin]

10 hours ago, Damocles Ti said:

Unfortunately these errors show nothing, and usually a CPU issue would put something in event logs half the time. I would also suspect it's RAM or memory controller. You can install something like HWinfo64, there's a WHEA listing there that loves to count up every time the memory controller is struggling to keep the FClock stable for your ram kit. Prime95 is not very good at testing memory but is excellent for testing Curve Optimizer stability (which we're not doing), and Memtest isn't good for testing something that's edge of stability either. I've found that Aida64 extreme's memory stress test works best at finding these barely stable memory overclocks.
 

• Try running HWinfo64 on a second monitor and check if the last row "WHEA" increases with your computer under load or gaming. If it does, your CPU's memory controller cant handle the RAM speeds. Might need SOC voltage tuning but that's a different rabbit hole
• Install Aida64 (trial version) and run the memory stress test under tools. It doesn't take very long, this baby usually finds errors within 10 minutes but I run it for an hour to be sure. If it quits, it means the memory OC isn't stable enough for regular use.

Yesterday, I had it running without any issues for about 10 hours. Today after 30 minutes I got freeze. Also had HWinfo running, but it froze aswell. 0 WHEA both yesterday (yesterday I had it for around 3 hours) and today.

 

Edit: it happened in following scenario - I had game and youtube running, went for toilet and resumed video, froze after couple of seconds.

[Damocles Ti]

Can you download and run Aida64, go into tools>system stability test>stress memory only>start, then wait an hour and check back? I have a suspicion it might be just the RAM overclock being unstable, and AIDA64 is really good at catching this

[Healadin]

On 6/30/2023 at 2:57 PM, Damocles Ti said:

Can you download and run Aida64, go into tools>system stability test>stress memory only>start, then wait an hour and check back? I have a suspicion it might be just the RAM overclock being unstable, and AIDA64 is really good at catching this

I did it yesterday, but firstly I switched back to Kingston RAM (the one certified for XMP, but not EXPO) because for those G.Skills I can still return them (and since they are not "working" out of box anyway, then I will just return them), also ran that AIDA64 memory stress test for almost 3 hours and there wasnt any problem, but after playing for another 2 hours, I got freeze once again (2 sticks of 16 GB Kingston FURY RENEGADE DDR5 RGB 6000 MHz CL32)

 

anyway I just tried lowering speed one profile lower (from 6000 MHz 32-38-38-80 1,35V to 5600 30-36-36-75 1,35V), and I guess I just have to wait now

[Healadin]

On 6/30/2023 at 2:57 PM, Damocles Ti said:

Can you download and run Aida64, go into tools>system stability test>stress memory only>start, then wait an hour and check back? I have a suspicion it might be just the RAM overclock being unstable, and AIDA64 is really good at catching this

aaand I already got like 3  freezes today, tried these 3 profiles so far

 

but noticed something interesting - that armoury crate I mentioned, that I am using for lights, when PC freezes, also the lights that go through aura freeze, havent noticed it until now, but maybe its somehow connected?

 

rest of the lights go through L-connect (lian-li software and they continue animating), only lights that are controlled by aura (AiO cooler and its fans that are mounted on top of case - that's why I havent noticed it until now) freeze

[Damocles Ti]

You can technically uninstall armory crate and then control RGB using a tool called OpenRGB, it uses less CPU than armory crate, and imo armory crate is almost a virus at this point. Reminds me I should clean install my windows because of the leftover undeletable junk Asus installs

[Healadin]

On 7/3/2023 at 7:14 PM, Damocles Ti said:

You can technically uninstall armory crate and then control RGB using a tool called OpenRGB, it uses less CPU than armory crate, and imo armory crate is almost a virus at this point. Reminds me I should clean install my windows because of the leftover undeletable junk Asus installs

don´t wanna jinx it, but I tried upgrading windows from 10 to 11 and it seems like it might have done the trick - Kingston RAM runs at 6000 MHz CL32 and whole weekend without a freeze

[Damocles Ti]

3 hours ago, Healadin said:

don´t wanna jinx it, but I tried upgrading windows from 10 to 11 and it seems like it might have done the trick - Kingston RAM runs at 6000 MHz CL32 and whole weekend without a freeze

I guess your OS had a few corrupted files then. Next time we could have started with running the command "sfc /scannow" in an elevated terminal to see what was up with windows. Either way, glad it seems like you fixed your issue with an "upgrade"