Question Possible problem with i9 9900k CPU

[ichaos42]

So, Ive been trying to track down why my computer will sometimes randomly shut down while in the middle of a game. Its only happened less than a handful of times in the last month but it got me wondering what was going on. I originally thought it was my Rog Strix 3080 ti GPU as it was the only new component i added recently (2 months ago).

In the course of google research it was suggested i download and run MemTest86 to check my Ram and CPU. Well, i ran the default test 3 times. each time took almost 5 hours and i get a FAIL everytime. But, when it says push any button to see summary, it is frozen on that screen. I am attaching some pics.
What i would like to find out is, is my cpu going bad? Because i have 3 months left on the warranty if it is.

I have not over-clocked this cpu. Never saw nor had a reason too. Imgur pics below:

Anyway, im open to ideas here if anyone has any?

https://imgur.com/q16AmWx

View: https://imgur.com/q16AmWx

https://imgur.com/gBiv4So

View: https://imgur.com/gBiv4So

https://imgur.com/RD8KgPa

View: https://imgur.com/RD8KgPa

https://imgur.com/0Yoxte8

View: https://imgur.com/0Yoxte8

https://imgur.com/JypCAIh

View: https://imgur.com/JypCAIh

 

[Lutfij]

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list them like so:
CPU:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:

Include the age of the PSU apart from it's make and model and the GPU you were on prior to the RTX3080Ti. Helps to know what you're using to cool the processor as well as the S version(not edition) if you're on Windows 10. BIOS version for your motherboard?

 

[Roland Of Gilead]

Hey there,

So, when your mem fails memtest, it typically point to one or both sticks of ram being faulty. Test your ram one stick at a time with memtest, and see if you can narrow it down to one stick.

Also, 3080ti's have issues with some PSU's (even very good ones). A bios update often solves these issues. The bios update will also ensure best compatibility with your ram.

 

[ichaos42]

Hey there,

So, when your mem fails memtest, it typically point to one or both sticks of ram being faulty. Test your ram one stick at a time with memtest, and see if you can narrow it down to one stick.

Also, 3080ti's have issues with some PSU's (even very good ones). A bios update often solves these issues. The bios update will also ensure best compatibility with your ram.

Oh, the memtest was failing because of my Ram? HUH.. I dont know why i thought it was my CPU then. i guess because i just saw the name of my cpu in the top of the screen. I will try that suggestion, testing one stick at a time.

As for the GPU i didnt realize you could do bios updates for those. I did and do make sure i keep my drivers updated. ill look into how i can get that taken care of. Thanks for the suggestions. It may take me a few days to post back but i will post the results once ive done them.

 

[ichaos42]

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list them like so:
CPU:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:

Include the age of the PSU apart from it's make and model and the GPU you were on prior to the RTX3080Ti. Helps to know what you're using to cool the processor as well as the S version(not edition) if you're on Windows 10. BIOS version for your motherboard?

Sorry, you are right and i actually meant to include all that information.
CPU: i9 9900k
CPU water cooler: Corsair H100l RGB PRO XT
Motherboard: Aorus Ultra Z390
Ram 32 GB Viper Steel ddr4 Extreme performance memory xmp enabled to run it at 4400
SSD 2T 980 Pro
SSD 1T 970
HDD 2T western Digital
GPU: Rog Strix 3080 Ti *previous GPU was EVGA FTW 1080
Psu: Corsair RM 850x *maybe 1.5- 2 yrs old
Chasis: Cooler Master CMSTORM Stryker
Monitor: Odyssey neo g9
OS: Windows 10

*nothing has been overclocked despite the functionality being there.

If there is anything else i should include, please, let me know.

 

[Roland Of Gilead]

Oh, the memtest was failing because of my Ram? HUH.. I dont know why i thought it was my CPU then. i guess because i just saw the name of my cpu in the top of the screen. I will try that suggestion, testing one stick at a time.

As for the GPU i didnt realize you could do bios updates for those. I did and do make sure i keep my drivers updated. ill look into how i can get that taken care of. Thanks for the suggestions. It may take me a few days to post back but i will post the results once ive done them.

So, I actually meant the bios update for your mobo (motherboard). Sorry if I didn't make that clear.

Yes, do let us know how you get on.

Also, just for informational purposes, yes, it is possible to upgrade a Vbios on a GPU. But not needed here. Not all GPU's support it. Some have to be force flashed (Flashing is the process of upgrading the bios).

 

[ichaos42]

Oh, the memtest was failing because of my Ram? HUH.. I dont know why i thought it was my CPU then. i guess because i just saw the name of my cpu in the top of the screen. I will try that suggestion, testing one stick at a time.

As for the GPU i didnt realize you could do bios updates for those. I did and do make sure i keep my drivers updated. ill look into how i can get that taken care of. Thanks for the suggestions. It may take me a few days to post back but i will post the results once ive done them.

So, i finally had time to get around to testing my ram one stick at a time. Each one failed. Then for some reason i had a thought, i went into my bios, disabled XMP and defaulted my ram speed to lowest setting 2133 Re-ran the test on a couple of the sticks and they all passed.
What gives? Why does it fail when it ran at the higher clock speed? Am i ok to just keep it running at the higher clock speed or should i keep it at the base 2133 speed?

 

[Phaaze88]

Just like core clock OCs, memory clock OCs can also degrade over time. The more aggressive it is, the sooner it usually happens.

The cpu's memory controller either:
A)needs more voltage to make that speed stable, only to fail again later on because of degradation, needing yet more voltage again to be stable. But you can't keep raising the voltage or you fry the memory controller.
B)is no longer capable of handling that speed regardless of voltage.

I looked up the kit and found the pdf for the 4400 kit: https://assets.website-files.com/5cdb2ee0b102f96c3906500f/6140d6fc7a6987508a104c33_PVS416G440C9K_Sku Sheet_050421.pdf
So profile 1 has a memory channel voltage of 1.45v. There really isn't any room to go higher than that.
There's a profile 2 of 4266mhz at 1.35v... is that stable for you?

 

[ichaos42]

Just like core clock OCs, memory clock OCs can also degrade over time. The more aggressive it is, the sooner it usually happens.

The cpu's memory controller either:
A)needs more voltage to make that speed stable, only to fail again later on because of degradation, needing yet more voltage again to be stable. But you can't keep raising the voltage or you fry the memory controller.
B)is no longer capable of handling that speed regardless of voltage.

I looked up the kit and found the pdf for the 4400 kit: https://assets.website-files.com/5cdb2ee0b102f96c3906500f/6140d6fc7a6987508a104c33_PVS416G440C9K_Sku Sheet_050421.pdf
So profile 1 has a memory channel voltage of 1.45v. There really isn't any room to go higher than that.
There's a profile 2 of 4266mhz at 1.35v... is that stable for you?

Profile one was unstable. kept rebooting pc. profile 2 was what ive been working on. From what you are telling me, its basically going to degrade over time no matter what. So, my question becomes.... how much benefit am i really getting from using profile 2?
As my first post stated, im not really having many problems with my pc. it just restarts sometimes in the middle of a game and not even that regularly.

 

[Phaaze88]

Yeah, once it starts, it's pretty much running on borrowed time... get it stable at a little lower settings and it works for awhile. Then it starts acting up again, and you gotta tune it down again... Rinse and repeat.
You can prolong it further by putting it back to stock/original speed - maybe you'll be ready to change it out before stock starts failing.


Hard to say about performance, as it's application dependent. It should still be better than nothing - if it's stable, I mean. I was curious if you had tested the second profile yet. Looks like you haven't finished.

 

[ichaos42]

Yeah, once it starts, it's pretty much running on borrowed time... get it stable at a little lower settings and it works for awhile. Then it starts acting up again, and you gotta tune it down again... Rinse and repeat.
You can prolong it further by putting it back to stock/original speed - maybe you'll be ready to change it out before stock starts failing.


Hard to say about performance, as it's application dependent. It should still be better than nothing - if it's stable, I mean. I was curious if you had tested the second profile yet. Looks like you haven't finished.

what do you mean test the second profile? It will only run on either disabled or the 2nd profile. it will not run on the 1st profile. *we are talking about xmp correct? in the bios?

 

[Phaaze88]

Is the 2nd profile memtest stable? I hadn't seen that, but I did see the profile 1 fails and the default 2133 passes.

 

[ichaos42]

Is the 2nd profile memtest stable? I hadn't seen that, but I did see the profile 1 fails and the default 2133 passes.

no. the only profile that passes is the default 2133. neither profile pass

 

[Psycotimis]

I have had a similar problem a year back; tracked the problem down to the PSU i upgraded it to a 1300W PSU never encountered the issue again. My suspicion is that the RTX 2080ti can draw more power than it says.