Question RAM errors when all four slots are populated ?

Dec 16, 2023
7
0
10
About a week ago, my pc suddenly refused to boot due to an error when loading the windows kernel (system32\ntoskrnl.exe). When I booted to my windows 10 installation media USB, I was met with a strange graphical error; I found a photo of another pc with same error on imgur. Link: View: https://imgur.com/Qk6N0


I figured I had either a faulty memory module or a faulty DIMM slot. After a bit of testing, I found that my motherboard would experience memory errors if all four DIMM slots were in use-- however all four of my memory modules and DIMM slots work perfectly, as I tested each individual module in each DIMM slot using memtest86. When all four memory modules are in the motherboard at once, memtest86 reaches an error count of 10000 and aborts the test. Three modules in any configuration pass the test with no errors. I tried running all four memory modules at lower speeds and lower voltages, but the errors persisted.

After removing two memory modules and successfully booting into windows, I used sfc /scannow and DISM /Online /Cleanup-Image /RestoreHealth to assure that windows was working fine. Furthermore, the audio coming from the motherboard's 3.5mm jack output is completely distorted: I only hear static. However, A cheap usb sound card provides working audio.

What I've tried so far:
1. Confirmed each individual memory module works
2. Confirmed each individual DIMM slot works
3. Running the ram at lower speeds & voltages
4. Updating the BIOS
5. Resetting BIOS via CMOS battery
6. Re-seating the CPU
7. Trying a Different CPU (Ryzen 5 2600x)
8. Verifying cable connections
9. Tested the PSU's voltages on all connectors (atx power, CPU power, pcie power)

To sum it up: using all four DIMMs on my mobo results memory errors and my onboard audio isn't working. Any help would be appreciated.

specs:
CPU: AMD Ryzen 7 5700x @ 3.4ghz
GPU: AMD Radeon 6700 XT
Motherboard: MSI x470 gaming plus
RAM: TEAMGROUP T-Force Vulcan Z DDR4 3200mhz 4x8gb
PSU: EVGA 750w N1 uncertified
Boot Drive: Kingston A400 120gb SATA SSD
 
Last edited:
officaly supported frequency for 5700x is 2666-2933MHz with four sticks, anything above is up to IMC quality (silicon loterry), mainboard trace layout, mainboard topology, mainboard layers and bios setting tweaks
 
officaly supported frequency for 5700x is 2666-2933MHz with four sticks, anything above is up to IMC quality (silicon loterry), mainboard trace layout, mainboard topology, mainboard layers and bios setting tweaks
Although I didn't mention trying those frequencies specifically, I have tried running at frequencies ranging from 2933-2133mhz, all of which produce the same memory error.
 
You bought this as a kit of 4 or two kits of 2?

This is scary with a 6700 XT. What's the exact model of the PSU?
I bought them as kits of two, they're from different batches but are otherwise identical.
I accidentally listed my psu as 650w, but it's actually a 750w n1 EVGA psu, which I purchased back im July of 2019.

What specific values for dram frequency and dram voltage did you try?

Set dram voltage to 1.35V and
set command rate to 2T.
I have tried running every available frequency in 2133-2933mhz range (2400mhz, 2733mhz, 2800mhz, etc.) I have tried each frequency at 1.35v and at 1.2v. My command rate is set to 2T, but I'm still getting errors.
 
This motherboard?

https://download.msi.com/archive/mnu_exe/mb/E7B79v3.0.pdf

In addition to not being matched RAM there is another possible issue.

Some motherboards require that the first physically installed RAM be placed in a specific slot.

If I found the applicable motherboard User Guide/Manual then that slot is DIMMA2 per numbered page 29 in the manual.

"Important. Always insert memory modules in the DIMMA2 slot first.

One of those steps that tends to be overlooked.....

Do verify that I found the applicable User Guide/Manual.
 
That means they're not identical and not guaranteed to work together properly.

Wow, that's really junk. You're risking your system with this unit.
The kits I bought are both rated for 3200hmz and have a cas latency of 16, the first kit I purchased in 2021, the other in 2023.I got all four to run stably at 3200mhz for a couple of months before then the errors began to appear.

Earlier I tested the psu with a multimeter and saw no abnormal voltages-- I'm still skeptical, and I've been planning on replacing that power supply for awhile now because of that.

Luckily, my local CTE center I'm enrolled at has an testbench and a plethora of spare hardware to work with, so I'll try my ram in a different motherboard, and vice versa.
 
This motherboard?

https://download.msi.com/archive/mnu_exe/mb/E7B79v3.0.pdf

In addition to not being matched RAM there is another possible issue.

Some motherboards require that the first physically installed RAM be placed in a specific slot.

If I found the applicable motherboard User Guide/Manual then that slot is DIMMA2 per numbered page 29 in the manual.

"Important. Always insert memory modules in the DIMMA2 slot first.

One of those steps that tends to be overlooked.....

Do verify that I found the applicable User Guide/Manual.
I have a physical copy of the manual, and I've already considered that. The PC posts fine with all 4 slots in use, but the moment the bootloader attempts to load the kernel into memory it halts because of an error, just as before.
 
Last edited:
"ini configuration, corrupt or missing system files, missing ntoskrnl.exe file, problematic Windows update, or missing driver for hard disk controllers. Also, as this error is related to memory, this may be caused by faulty RAM.Sep 20, 2023"

Source:

ntoskrnl error...

Memory remains suspect because the RAMs are not a matched set.

Try setting boot to verbose so the boot process will display more about what is happening during boot.

Slows the boot process down some but that is a necessary trade-off for troubleshooting purposes.

= = = =

Other suspects:

Source for Windows OS....

Corrupted Installation media....

Failed or corrupted update....

Search for and find other links regarding ntoskrnl,.exe. Read and modify search criteria as you learn more and can match symptoms to the problem halts.
 
"ini configuration, corrupt or missing system files, missing ntoskrnl.exe file, problematic Windows update, or missing driver for hard disk controllers. Also, as this error is related to memory, this may be caused by faulty RAM.Sep 20, 2023"

Source:

ntoskrnl error...

Memory remains suspect because the RAMs are not a matched set.

Try setting boot to verbose so the boot process will display more about what is happening during boot.

Slows the boot process down some but that is a necessary trade-off for troubleshooting purposes.

= = = =

Other suspects:

Source for Windows OS....

Corrupted Installation media....

Failed or corrupted update....

Search for and find other links regarding ntoskrnl,.exe. Read and modify search criteria as you learn more and can match symptoms to the problem halts.
I'll try enabling verbose boot and analyze the feedback I get. I'm going to test my ram tomorrow in another working PC, hopefully. Also worth noting: I attempted to boot to a Fedora 39 workstation live USB, and although I was able to select the OS in the grub bootloader, the Linux kernel never loaded successfully. I don't believe it's windows issue.