hardware issue Crashing...A LOT

[H60MTI]

Yeah, that's all stuff that I've run today. DCS obviously. I was running Cinebench and FurMark earlier to try and see if my rig was stable. Both of those are stress testers. FurMark is a graphics stress test and Cinebench is a CPU stress test. Ghost Recon: Breakpoint is another game that I play. So, if I'm seeing hardware and not handling memory properly, I wonder if a RAM stick isn't seated right or has going bad....

Just ran those 2 prompts. I'm going to shut down actually and reseat all my RAM for <profanity>s and giggles.

[H60MTI]

WTF? I can't curse anymore?! That sucks.... Thanks ED... We're back. Updated DCS. Ran rebuild and cleanup. Ran dism and scannow prompts. Shut down and reseated my RAM. Let's see what happens...

[H60MTI]

And crash....Oh my God this is annoying.

DxDiag.txt

[H60MTI]

Well, we're past DCS and games crashing. Definitely a rig issue and I don't know what. I'm getting looping BSOD with restarts or shut downs. Not even time to troubleshoot cause it BSOD's. I think I'll tear the rig apart and rebuild it. This is ridiculous. 

Thanks for the help guys, I do appreciate it. Cross your fingers and pray. 

[HansPeter1981]

Hello,

When I upgraded to a 7900xtx from a 6800xt I expirienced similar issues... was so sure that my psu was enough and couldnt see a problem with Furmark stress testing.

Eventually I installed HWmonitors newest version and now I was able to see the short term peak power draw of my GPU turned out it was well over 440 Watts.

Bought a new 1000w PSu... no more problems.

 

With those shutdowns along game a safety feature of my dcs Sata drive so eventually it wouldnt show after restart... was s9me sort of trick to get it back alive plugging it in at the right moment.

 

Long story short it was those peaks that would cripple my 650 Watt pwr sply. Fumark at a 100% gpu would never create this. Newer games like Breakpoint would...

[sleighzy]

Really sorry to hear that mate.

From that last DxDiag log:

• Couple more of the LiveKernelEvent 141 errors
• Some blue screen of deaths which you've noted (don't have a system code, or not seeing it properly, for cause of these)
• ASUS Armoury software (ArmourySocketServer.exe)
• Crashes with the Windows login screen (LogonUI.exe)
• Another one for DCS with the same memory problem

If you do manage to get into your system, even if starting up in safe mode, there are some items you could check here, one of them being a memory test.

Don't know if you're getting any system codes etc. on your blue screen which may indicate more around the cause.

https://support.microsoft.com/en-us/windows/resolving-blue-screen-errors-in-windows-60b01860-58f2-be66-7516-5c45a66ae3c6

Edited November 5, 2024 by sleighzy

[H60MTI]

I'm back-ish. I'm thinking a bad RAM stick. Right now I'm running on one RAM stick, so 16GB of DDR4 GSkill Trident Z RGB. Good stuff. I typically run 4 sticks of said RAM. Before I went offline I was having constant/looping BSOD's. So, I'm running one stick at a time and seeing if I find a bad stick that won't boot Windows.

[H60MTI]

So, it has to be a bad RAM stick or a bad channel. I tried each stick separately and computer booted just fine. At the moment I'm in DCS and it's kicking with 2 sticks in dual channel mode. When I get some more free time, I'll try all 4 sticks again and see what happens. PITA!!!

[sleighzy]

Oh well, good progress anyway, at least you’ve narrowed down the cause. Small consolation, but you get to play DCS without it crashing on you

You could also add a larger page file to help with the loss of RAM. If you do decide to then just add a custom page file on your fastest SSD with initial and max values of 32768, click set, apply, and restart your PC. Make sure you only have one page file, eg remove any system managed ones. Might help with performance where needed until you get your full physical 64gb ram back again. 

[sleighzy]

One thing that sprung to mind is you mentioned you have 4 sticks of RAM. You could check if you have an XMP profile set in your BIOS for your RAM, and turn it off if so. Four sticks may be more susceptible to timing issues and crashes than two sticks when using some XMP profiles.  Worth a shot to rule it out. 

[H60MTI]

I'll definitely give it a shot. I do have XMP enable. With a Ryzen CPU of course I have XMP enabled Lol I wasn't really able to get into BIOS before cause of the looping BSOD's. 

[H60MTI]

Update...Was running on 2 sticks of RAM, 32GB total, since last post. I've got all 4 sticks back in to try and narrow down the bad stick and or bad slots in my motherboard. So far so good, some how. Was running Ghost Recon: Breakpoint for a bit and I'm swapping over to DCS as we speak. We'll see what happens.

[silverdevil]

2 hours ago, H60MTI said:

Update...Was running on 2 sticks of RAM, 32GB total, since last post. I've got all 4 sticks back in to try and narrow down the bad stick and or bad slots in my motherboard. So far so good, some how. Was running Ghost Recon: Breakpoint for a bit and I'm swapping over to DCS as we speak. We'll see what happens.

i did not go through all the posts on this thread after i posted. forgive me if it has been suggested. sometimes RAM could be simply be reseated to correct issues. 

[H60MTI]

I'm thinking that's all it was. I put all my sticks, 64GB worth, and it's running fine now. Was playing Ghost Recon: Breakpoint a bit earlier and now I'm in DCS with no issues. Seems like a stick just needed to be reseated for reason. Don't know why after a couple of years in it. Beats me but she's running like she usually does.

[Akula]

I had to disable hyperthreading for my 14900k and it solved literally every crash I’ve had. It knocked me down to 16 physical cores so it’s not too bad. I can’t even tell the difference in performance and I have actually been TRYING to trigger a crash. All mods work. Spamming the F2 map only caused a crash on MP but that’s because i was forcing the video card to create and recreate images faster than i was hitting F2 which it could not do, hence the crash. That’s it. Play the game normally and I don’t have a single issue. 

[silverdevil]

11 hours ago, Akula said:

I had to disable hyperthreading for my 14900k and it solved literally every crash I’ve had. It knocked me down to 16 physical cores so it’s not too bad. I can’t even tell the difference in performance and I have actually been TRYING to trigger a crash. All mods work. Spamming the F2 map only caused a crash on MP but that’s because i was forcing the video card to create and recreate images faster than i was hitting F2 which it could not do, hence the crash. That’s it. Play the game normally and I don’t have a single issue. 

i know this CPU has had its share of problems. i read an article yesterday that may surprise some users.

https://www.xda-developers.com/intel-core-i9-14900k-deteriorated/

[Akula]

5 hours ago, silverdevil said:

i know this CPU has had its share of problems. i read an article yesterday that may surprise some users.

https://www.xda-developers.com/intel-core-i9-14900k-deteriorated/

That’s a good article and I agree, this chip sucks for stability. I hate it but I’m stuck with it. I don’t know how so many people are buying 2 or three chips and testing. I guess they have the cash?

But seriously, to put that article into perspective it was released on Nov 11 with the title:

- “After only one year of use, my Intel Core i9-14900K has deteriorated to the point that it's unusable”

So like many who bought this chip in Nov 2023, they were under the Intel-fed false impression that the 14gen chips could take the heat of overclocking so he probably overclocked like crazy trying different methods which degraded the chip pretty fast considering there was no microcode controlling the voltage spikes and proportionately matching power to performance. In other words, Intel was allowing even mild overclocking to get 1.3+ volts and spiking at 1.55 and beyond (this is November 2023 we are talking about, not today)

I got mine in Feb 2024 knowing full well that the first thing I had to do was undervolt it. Honestly, I probably ran this thing at 1.4-1.5 volts once or twice during initial testing but by Feb 2024 there was already a lot of usefull info by other overclockers on how to get the voltages between 1.2 and 1.3 which is the range I’ve suck with once I learned how to OC properly. 
 

That said, the jury is still out on if I have degraded this chip towards a fast downward spiral as it seems that once you have defeated these chips past a certain point you can hardly hold a good clock speed and need to keep volting UP to maintain it this killing off the chip even more. I could be wrong but in my research that seems to be what is hapenning. 
 

if my chip dies in a year I can still claim it on warranty. No big deal. 
 

Had the 9800X3D been available back in February I would have certainly gone with that setup knowing what I know now. 
 

Right now I’m running everything at 5.4 @1.2 volts under load with no voltage spikes. That’s pretty damned good!
   Windows is running like a champ with hyperthreading disabled and like I said; With 16 cores left to play with I’m really not seeing or feeling a drop in performance.

[Akula]

On 11/5/2024 at 7:36 PM, H60MTI said:

So, it has to be a bad RAM stick or a bad channel. I tried each stick separately and computer booted just fine. At the moment I'm in DCS and it's kicking with 2 sticks in dual channel mode. When I get some more free time, I'll try all 4 sticks again and see what happens. PITA!!!

Hey try this trick. It’s an oldie but it may help.

your ram sticks are all numbered in order as per the PHYSICAL stickers on each of the ram sticks. When you look at the stickers you’re going to see a bunch of numbers and letters. Don’t worry about that. Just look for similarities where the numbers hint that one stick is the first and the second stick is the second. If you buy a two stick kit you’re going to see this easily. Same goes for buying a 4 stick kit. 
Now, if you’re like me and bought 2 sticks and then later you bought another 2 stick kit of the exact same make and model memory then you need to arrange them properly. The first set needs to go into your primary dimm slots in order (first stick goes in the primary slot and second in the secondary). Then you will want to place your second kit in the same order but in the third and fourth fill slot respectively. As long as install the sticks in their proper order according the info on the stickers it will help. How much it helps depends on your system. Everyone’s results will vary but either way it’s gonna be better than just installing them willy nilly and hoping for the best. 

Edited November 17, 2024 by Akula