Q-Code 54 on all RAM kits, only fixed after BIOS downgrade

[HuMeN]

Hi ASUS team and community,

If this issue has already been discussed here, I apologise – I wasn’t able to find a complete match. I’d appreciate a link if someone has already shared a solution.
Still, I wanted to report my case because it’s been persistent across multiple RAM kits and may help others—or even contribute to improving future BIOS versions.

System:

• CPU: Intel Core i9-13900K (watercooled)

• GPU: EVGA NVIDIA RTX 3080 Ti FTW3 Ultra (also tested without GPU)

• Board: ASUS ROG Maximus Z790 Extreme

• BIOS: 2801 (before), now running 2202

• RAM Kits tested:

  1. G.SKILL Trident Z5 RGB 7600 CL36
     (Model: F5-7600J3646G16GX2-TZ5RK, 2×16 GB)

  2. G.SKILL Trident Z5 RGB 7800 CL36
     (Model: F5-78800J3646H-16GX2-TZ5RK, 2×16 GB)

  3. Kingston Fury Renegade 6000 CL32
     (Model: KF560C32RSK2-96, 2×48 GB = 96 GB)

Problem Overview (on BIOS 2801):

• The system consistently failed with Q-Code 54 (Memory initialization error) on all RAM kits.

• I tested:

  • All recommended slot configurations (A2+B2, B2 only)

  • All other valid slot combinations, including fallback patterns like A1+B1 or A1 only

  • Default settings, no XMP

  • XMP I and XMP II

  • Manual speeds (6000 / 5600 / 4800 MHz)

→ No combination worked – the system always ended with Q-Code 54.

I also tested without the dedicated GPU to see if the iGPU from the i9-13900K would work.
The system reached Q-Code 30, indicating a completed POST, but no display output from the iGPU.
This only changed after downgrading the BIOS.

What fixed it:

I downgraded to BIOS 2202 using EZ Flash 3.
After the downgrade:

• The system booted successfully with both 48 GB Kingston modules

• Q-Code 54 disappeared completely

• Display output from the iGPU was restored immediately

• Boot times were back to normal

• JEDEC 4800 MHz (Auto settings) ran fully stable

Post-downgrade confirmation:

After the downgrade, I:

• Retested all three RAM kits again to rule out module errors

• Reinstalled my dedicated GPU (RTX 3080 Ti)
  → System worked flawlessly with iGPU and dGPU
  → No more error codes, no black screens

Conclusion:

BIOS 2801 introduces serious memory training issues that affect multiple high-end and high-capacity DDR5 kits – including my previously stable G.SKILL that had been running flawlessly for nearly 3 years on this platform before BIOS 2801 introduced the issue — so this is not a matter of incompatible hardware, but likely a regression in BIOS behaviour.
Only a downgrade to BIOS 2202 restored full system functionality.

Suggestion to ASUS:

Please review memory training logic introduced in BIOS 2801. These BIOS versions appear to be overly aggressive or insufficiently compatible with RAM configurations that previously worked flawlessly – even those listed on the official QVL. In my case, all three tested kits failed under BIOS 2801.

Users should not be penalised for keeping their BIOS up to date, especially if it results in loss of basic functionality like POST failure or no display output.

Thanks in advance for looking into this. I hope this helps others facing similar problems.

[MasklessFate]

> BIOS 2801 introduces serious memory training issues 

Do you have any AMD system you can try this with? I haven't updated my bios in a while, so maybe it also happens on AMD. If it does try enabling Memory context restore by searching for it and then enabling both of the options.

[Vynra]

memory context restore is literally only for AMD. How is that going to help on an intel system?

Also to OP. Have you tried upgrading back to 2802? is it possible the bios update got corrupted somehow?

[MasklessFate]

OP mentioned memory training, that is why I asked if they had an AMD board, so they could see if it was really memory training or not...

[Silent_Scone]

New ME Firmware and u-code can impact overclocking. I would also advise you to remove any contact frame if using one as depending on the pressure mounting it can impede OC range.