Issue with PC showing a splash screen on the motherboard during reboot and occasionally during a cold start.
Issue with PC showing a splash screen on the motherboard during reboot and occasionally during a cold start.
Recent updates have shown a significant rise in my system freezing on the motherboard splash screen during startup. At first, it was occasional and only occurred during reboots; turning the machine off and on resolved the problem. Recently, I spent nearly half an hour repeatedly powering the PC to reach the desktop state, which is quite different from the brief freezes during reboots.
Since getting the computer, I haven’t made many changes. I’ve updated the NVIDIA drivers (currently at version 581.08) and turned on Maximum Performance in the NVIDIA Control Panel as well as Windows Power Plan settings. I also activated EXPO1 in UEFI. I don’t overclock either the GPU or CPU. Although game performance isn’t great for this configuration, it hasn’t been noticeably impacted by whatever’s causing the freezing.
Before this problem, I tried enabling PBO from Auto to Enabled, but the system would hang at the splash screen. I disabled it afterward. It’s worth mentioning that I enabled EXPO1 a few days prior to attempting to switch PBO, which coincided with the start of this issue. I’m not claiming the memory profile is the cause, but everything seemed to change around the same time.
Temperatures look normal. I track them using NZXT Cam; my CPU stays between 70-80°C under heavy use (like Madden/Tarkov), and the GPU around 60-70°C. Slightly warmer than ideal for the processor (around 45-50°C), though some say the 7800X3D handles heat better.
I haven’t performed any serious troubleshooting yet, as I’m unsure what might be responsible. I considered booting into UEFI and disabling EXPO1 to test, but given my recent experience, I thought it wise to research first before rebooting. I also didn’t attempt to re-seat components since they didn’t appear loose when the PC was delivered in May and hasn’t been moved since.
Also, on my previous Intel system, I had XMP1 enabled without issues, so I assumed EXPO1 would be the right setting here too. EXPO2 seemed to be running around 5k rather than the expected 6,000 speeds when I first enabled it.
Please let me know if you need more details to help figure out this problem.
MSI MAG X870 Tomahawk
UEFI v1A44 from 4/24/25
https://www.msi.com/Motherboard/MAG-X870...FI/support
The current BIOS version is 7E51v1A67. It might be useful to manually install the newest chipset drivers from AMD's support page using an elevated command, such as right-clicking the installer and selecting Run as Administrator, then updating the BIOS for your motherboard.
Check for any motherboard warning indicators during startup. Ensure the system isn't attempting memory initialization at boot.
Here are some notes based on your input:
- I wanted to start by mentioning that I haven't kept detailed records of everything, so much of this comes from vague memories.
- This has been a recurring problem for months, and I eventually had to accept it.
- Recently, with the launch of Arc Raiders, I preordered their AREX0024 error code, which often disconnects me, requiring only a system reboot to resolve.
- The daily reboot issue is consistent: the screen turns dark briefly, shows the splash, spins the loading circle, then freezes for 20-30 seconds before turning off. The only solution is to power down the system.
- Once cold started, it usually boots normally, but sometimes it loops back into the hang.
- I should note that aside from this, the system runs well once installed.
- The PC arrived in mid-May; early on I didn’t notice major problems, but I did encounter boot issues after installing EXPO and updating to UEFI version A67.
- Recent tests suggest EXPO and Memory Context Restore (MCR) might be causing the instability.
- EXPO1 at 6,000 MHz with MCR ON gives intermittent POST; EXPO1 + MCR OFF shows no POST at all.
- I’ve tried turning off Power Down Enable, as it seemed unnecessary based on research, but the tests were limited in scope.
- Clearing CMOS twice helped for a short time, and another reset resolved the issue temporarily.
- From what I found online, this motherboard may have reliability concerns, possibly due to weak IMC or poor DRAM quality.
- I reached out to NZXT for advice, but they recommended returning the product, with a long wait time due to distance.
At this stage it appears there is another BIOS release available;
https://www.msi.com/Motherboard/MAG-X870...FI/support
Version 7E51v1A69
It seems my IMC might be the problem.
If you suspect the IMC is the cause, consider replacing the CPU and check if the same issue appears with the other components.
The DRAM package I received is quite poor.
Should you believe the RAM package is responsible, try transferring it to another functional system (with the correct BIOS) and observe if the problem resolves.
If neither the RAM nor the CPU shows the same issue as your prebuilt unit, the fault likely lies with the motherboard—consider returning it or replacing it.
Right now, you purchased a brand new machine intended to work immediately. Investigating a system is essentially squandering the support you paid for while investing in a ready-made solution. In line with NZXT’s approach, I’d advise avoiding them and constructing a custom build instead.
Please keep it simple—stay in one thread unless you’re ready for chaos.
Unfortunately, I can't test these components in isolation. Either I notice something I missed or I move forward with the RMA process that could take months. I’d rather avoid the latter.
I noticed the new BIOS release, but it doesn’t mention any memory compatibility fixes or updates. I believe the AGESA update affects EXPO, though I’m not entirely sure. I don’t have much background on this topic.
And yeah, sorry for essentially rewriting this post, but I wasn’t clear on the guidelines for posting my own thread since it was from two months ago.
i think there might be a small mix-up in the model number on the support list—it seems like nzxt was assembled by teamgroup and they have a wide range of dimms. it doesn’t really matter or make a difference, right? it should work fine. it sounds like this is the team’s best guess, so you could try compatible tested dimms or wait about two months for a turnaround. maybe you can check at a microcenter or shop before committing if you want to avoid any issues.
CPU World indicates the 7800X3D is configured for DDR5-5200, yet it operates at speeds of 6,000MT/s and 5,600MT/s.
Memory controller details: two channels available, supported memory type DDR5-5200, ECC enabled.
It's possible you didn't get a favorable outcome in the silicon selection process.
Exceeding the manufacturer's suggested clock speeds can lead to frustration, particularly if your RAM isn't reliable. A rate of 6,000MT/s surpasses the official AMD limit of 5,200MT/s. Results vary—sometimes success, sometimes not.
Before attempting XMP/EXPO/DOCP overclocking, I perform multiple rounds of MemTest86 (several hours each), only accepting a stable result after zero errors.
Currently, I still use the 7950X at its stock DDR5-4800 configuration. For long video renders that don't gain much from faster RAM, stability is more important than minor speed gains.
For gamers, boosting RAM speed might help FPS, but only if the system remains functional.
Turn off EXPO, run MemTest86 to detect issues, then manually adjust RAM speed to 5,200MT/s and complete a full test. If stable, try 5,400MT/s, repeat the process, increasing by 200MHz increments until errors appear, then lower slightly.
https://www.memtest86.com/
Thank you for the input. It seems the MSI QVL for this board might not be complete yet, and it's possible the model number hasn't been added yet. The system works fine after booting into the desktop, which makes me think the issue could be related to firmware or BIOS. I'm encountering mixed information while researching; some suggest ICM/DRAM problems but no performance issues. Since I don't have another AM5 board for comparison, I'm trying to pinpoint symptoms carefully before assuming a defect or facing a long RMA process.
I know the AMD JEDEC spec is 5200, but that's only the minimum requirement, right? AMD suggests 6000 as the ideal range. My concern isn't about performance drops. I haven’t seen WHEA errors, BSODs, or random crashes, and no data corruption has occurred. I doubt MemTest would reveal anything. Once the system starts up, everything works fine. It’s a bit disappointing given the specs, but it’s acceptable.
It looks like I’m stuck in a memory loop and I can’t tell if it’s hardware-related (like my IMC or DRAM) or if the BIOS version is faulty. The X870 is new (September 2024), and early AMD builds are known for this kind of issue.