Memory Problems on ROG MAXIMUS Z790 HERO

[edup_pt]

I've bought a ROG MAXIMUS Z790 HERO with the following configuration:

- CORSAIR Vengeance DDR5 RAM 128 GB (4 x 32 GB) 5600 MHz CL40 Intel XMP iCUE (CMK128GX5M4B5600C40)

- Processor Intel i9-14900K LGA 1700
- SSD Samsung 990 PRO 2TB PCIe 4.0 NVMe M.2 (2280) (MZ-V9P2T0BW)
- HP NVIDIA RTX A5000 24GB 4DP GFX

At the very beginning I had problems with the 128Gb of ram which after the bios update disappear. Though, after a while I started to understand that the computer was freezing by himself of after a while of training models. The errors are completely randoms. I know that is related with memory because i run some tests with different memory speed and was observing errors with the speed increase.

Currently, I have this memory configuration:

# dmidecode 3.3

Getting SMBIOS data from sysfs.

SMBIOS 3.5.0 present.

Handle 0x000D, DMI type 17, 92 bytes

Memory Device

Array Handle: 0x000C

Error Information Handle: Not Provided

Total Width: 64 bits

Data Width: 64 bits

Size: 32 GB

Form Factor: DIMM

Set: None

Locator: Controller0-DIMM0

Bank Locator: BANK 0

Type: DDR5

Type Detail: Synchronous

Speed: 4200 MT/s

Manufacturer: Corsair

Serial Number: 00000000

Asset Tag: 9876543210

Part Number: CMK128GX5M4B5600C40

Rank: 2

Configured Memory Speed: 4200 MT/s

Minimum Voltage: 1.1 V

Maximum Voltage: 1.25 V

Configured Voltage: 1.1 V

Memory Technology: DRAM

Memory Operating Mode Capability: Volatile memory

Firmware Version: Not Specified

Module Manufacturer ID: Bank 3, Hex 0x9E

Module Product ID: Unknown

Memory Subsystem Controller Manufacturer ID: Unknown

Memory Subsystem Controller Product ID: Unknown

Non-Volatile Size: None

Volatile Size: 32 GB

Cache Size: None

Logical Size: None

Handle 0x000E, DMI type 17, 92 bytes

Memory Device

Array Handle: 0x000C

Error Information Handle: Not Provided

Total Width: 64 bits

Data Width: 64 bits

Size: 32 GB

Form Factor: DIMM

Set: None

Locator: Controller0-DIMM1

Bank Locator: BANK 0

Type: DDR5

Type Detail: Synchronous

Speed: 4200 MT/s

Manufacturer: Corsair

Serial Number: 00000000

Asset Tag: 9876543210

Part Number: CMK128GX5M4B5600C40

Rank: 2

Configured Memory Speed: 4200 MT/s

Minimum Voltage: 1.1 V

Maximum Voltage: 1.25 V

Configured Voltage: 1.1 V

Memory Technology: DRAM

Memory Operating Mode Capability: Volatile memory

Firmware Version: Not Specified

Module Manufacturer ID: Bank 3, Hex 0x9E

Module Product ID: Unknown

Memory Subsystem Controller Manufacturer ID: Unknown

Memory Subsystem Controller Product ID: Unknown

Non-Volatile Size: None

Volatile Size: 32 GB

Cache Size: None

Logical Size: None

Handle 0x000F, DMI type 17, 92 bytes

Memory Device

Array Handle: 0x000C

Error Information Handle: Not Provided

Total Width: 64 bits

Data Width: 64 bits

Size: 32 GB

Form Factor: DIMM

Set: None

Locator: Controller1-DIMM0

Bank Locator: BANK 0

Type: DDR5

Type Detail: Synchronous

Speed: 4200 MT/s

Manufacturer: Corsair

Serial Number: 00000000

Asset Tag: 9876543210

Part Number: CMK128GX5M4B5600C40

Rank: 2

Configured Memory Speed: 4200 MT/s

Minimum Voltage: 1.1 V

Maximum Voltage: 1.25 V

Configured Voltage: 1.1 V

Memory Technology: DRAM

Memory Operating Mode Capability: Volatile memory

Firmware Version: Not Specified

Module Manufacturer ID: Bank 3, Hex 0x9E

Module Product ID: Unknown

Memory Subsystem Controller Manufacturer ID: Unknown

Memory Subsystem Controller Product ID: Unknown

Non-Volatile Size: None

Volatile Size: 32 GB

Cache Size: None

Logical Size: None

Handle 0x0010, DMI type 17, 92 bytes

Memory Device

Array Handle: 0x000C

Error Information Handle: Not Provided

Total Width: 64 bits

Data Width: 64 bits

Size: 32 GB

Form Factor: DIMM

Set: None

Locator: Controller1-DIMM1

Bank Locator: BANK 0

Type: DDR5

Type Detail: Synchronous

Speed: 4200 MT/s

Manufacturer: Corsair

Serial Number: 00000000

Asset Tag: 9876543210

Part Number: CMK128GX5M4B5600C40

Rank: 2

Configured Memory Speed: 4200 MT/s

Minimum Voltage: 1.1 V

Maximum Voltage: 1.25 V

Configured Voltage: 1.1 V

Memory Technology: DRAM

Memory Operating Mode Capability: Volatile memory

Firmware Version: Not Specified

Module Manufacturer ID: Bank 3, Hex 0x9E

Module Product ID: Unknown

Memory Subsystem Controller Manufacturer ID: Unknown

Memory Subsystem Controller Product ID: Unknown

Non-Volatile Size: None

Volatile Size: 32 GB

Cache Size: None

Logical Size: None

I also tested with this last firmware update (beta) 2602. I'm running "Ubuntu 22.04.4 LTS".

The computer is currently not useful to me because of this problem. I already removed the cpu and put it back on motherboard. My hardware is completly unstable. 

Can you give me any direction? Thanks in advance.

[achugh]

Hi @edup_pt typically in a situation like this you should can try the following approach.

1. Turn off XMP and use the default JEDEC speed which will be around 4800MT/s, now run your memory tests to confirm everything is stable.
2. Assuming you are stable then enable XMP II, run your stability tests again. You should be able to get closer to 5200 MT/s 
3. Assuming you are stable then enable XMP I or XMP Tweaked (if shown for your system), run your stability tests again. You should be able to get closer to 5600 MT/s 

As a reference you can read more about memory overclocking at https://rog-forum.asus.com/t5/technologies-explained/memory-kits-overclocking-and-what-you-may-not-k...

If you are someone who understands better from videos, checkout https://www.youtube.com/watch?v=wzVZgTP2204 

I hope this information helps you resolve your issue.

[edup_pt]

Thanks so much for your reply.
I don't have an option to disable XMP.   I have an option to change AI Overclock Tuner to manual but then 2 values appear BCLK FREQ and PCIE FREQ (they are both 100 by default). I can change the memory speed independent on the value of AI Overclock Tuner. I tested with 4800MHZ and I still have issues. I'm lowering to 4600MHZ (testing now). Anyway this probably means that the board have some kind of compatibility problem with this memory, right?! Not sure what I can do more.

[Silent_Scone]

Hi @edup_pt 

Overclocking this much memory often comes with some concessions and not all CPUs will be able to achieve the same frequency beyond stock.

Please see the below chart for what's considered an overclock depending on the memory configuration on 14th gen CPUs.

1. Two DIMMs per Channel on a Two Slot per Channel board using dual rank DIMMs results in a maximum supported speed of 4400MT/s. Anything above this can be considered an overclock whilst in this configuration.

2. XMP overclocking can be disabled by either clearing the CMOS via the motherboard's rear IO button or entering the UEFI and going to Extreme Tweaker > Setting AI Overclock Tuner to Auto.

[edup_pt]

Thanks for the reply. It makes sense.

I tested with 4800MT/s, but still with issues if i start to work with more memory.

Trying now with the 4400MT/s standard speed.

[edup_pt]

I tested with  4400MT/s but still no luck. I have 4 slots with 32GB  - CORSAIR Vengeance DDR5 RAM 128 GB (4 x 32 GB) 5600 MHz CL40 Intel XMP iCUE (CMK128GX5M4B5600C40) .

I'm tested with 4200MT/s and the system stalls...... I'm starting to think that there's any hardware problem, even that i already tested the memory before.

Should i try with less speed? Thanks so much

[achugh]

Hi @edup_pt have you tried to update your BIOS to the latest version recently released followed by CLEAR CMOS then F1 into BIOS and F5 to load latest defaults, finally F10 to SAVE and EXIT.

Does this default configuration work for you as a stable state?

[edup_pt]

Hi @achugh , I've updated bios to latest version but on this last one i don't remember if i loaded the defaults. It was working with 128gb and still unstable. I've removed the 2 slots of 32gb (64gb) of memory and currently it's working stable with the other 2 slots of 32gb (64gb). I will try later what you're suggesting, but this is not the first time that this motherboard doesn't work well with 128gb of memory. It was never stable with that amount of memory.

[achugh]

Hi @edup_pt from your board's support page, your vendor (for 4 x 32 GB configuration) and specifically this memory kit you are using is not listed on the QVL page. I have typically found that in these scenarios it gets difficult to get things working.

I picked my memory 192 Gb (4 sticks of 48 GB each) as a kit from the QVL of my board ROG MAXIMUS Z790 DARK HERO and I am able to run my kit at its rated speed of 5200 MT/s without any issues.

My system details are in https://rog-forum.asus.com/t5/gaming-motherboards/can-asus-support-fix-realtek-audio-drivers-with-bo... post.

Depending on your level of knowledge or the desire to find a solution, you can manually tweak your memory timings if the provided XMP profiles are not working. Or you can sell this kit and switch to one of the supported kits from the QVL (all of them are from Kingston).

Good luck!

[edup_pt]

You're right. Probably this new drivers messed up with this kind of memory, but they also don't say that this memory is compatible. Going to sell my 128GB of ram since neither with 64gb it works ok with the default settings.

Here are some pictures of the errors on journalctl.  Maybe it helps who write the drivers.