System fails to start or post and does not boot when XMP is activated following RAM installation
System fails to start or post and does not boot when XMP is activated following RAM installation
Hello, I recently upgraded my system by adding an extra 16GB of RAM alongside my existing 16GB. After trying to boot with the full 32GB, the setup didn’t start properly and displayed a three-beep error. Once I swapped out the modules, the system powered on successfully, but XMP remained disabled. I re-enabled it after some adjustments, but it still showed the same error code. Eventually, I reset everything, left only one module active, restored BIOS access, and kept XMP off after booting. Some guidance suggests this issue might relate to BIOS clock speed settings (set to 'auto'), but I’m unsure what to do next. My hardware specs are listed below:
- Motherboard: GIGABYTE AM4 B550 AORUS Elite V2 (Rev 1.2) - BIOS revision 'FEd'
- CPU: Ryzen 5 5600G
- RAM: 4x G-SKILL Ripjaws V DDR4-3600 CL18
- PSU: Corsair CX550 80+Bronze
I’m new to BIOS tweaks, so I haven’t tried other fixes beyond what’s mentioned. Please let me know if keeping XMP disabled is the right approach.
Probably. Even with identical model numbers, these sticks are likely distinct. Memory makers often adjust the ICs used in kits based on cost fluctuations, making it probable they differ here. Various ICs require unique training processes, which can cause compatibility problems, particularly at faster speeds. To improve performance, I’d start by placing the older kit in slots 1 and 2, newer in 3 and 4. This separates them by memory channel, allowing individual training per channel—boosting compatibility, though AMD setups generally see less benefit than Intel. While the sticks are available, check the top-right label for the coded IC; compare the last and third-to-last characters (e.g., 04213X8810B vs 04266X8811B) to confirm they match. Once verified, manually input XMP settings—adjust voltage and timing—and gradually increase frequency from the stock 2133MT/s until it fails, then perform a stress test at that peak speed.
I hadn’t considered how much time six months could change things. To be sure, here’s what I’d do next: First, if I can’t restore everything with all four drives, the safest step is to clear everything and try booting with just one drive. Second, once I have all four installed, I’ve already seen internet problems—slow speeds and odd HTTPS errors. It seems more likely tied to RAM or other hardware factors rather than just the drives themselves.
Reset the system by clearing the CMOS. You might find a connection point where you can short with a screwdriver or reset the case button to it. Alternatively, remove the battery on the motherboard for a short while when the power supply is off. Unstable memory can lead to problems like this, so consider doing a stress test afterward. It seems unrelated though.
You were correct about the differences in the sticks. One group uses IC 04320X8811E while another employs 04213X8820D. It's interesting that both labels suggest a short gap—just a month between October 2023 and September 2023. This implies these sets might not be compatible. I plan to test them and run a stress test to catch any issues. You're likely right too; the connection problem continues even after installing just the old pair. It probably points to a problem with the garbage dongle, possibly due to firmware errors from previous reboots.
For additional details here: BIOS documentation indicates one board uses Hynix controllers while another employs Samsung controllers. Got some success boosting frequencies beyond the default, but at around 2333 MHz it failed to boot. I had to remove the CMOS to restore the BIOS settings. Because I’m not confident enough to adjust memory speeds myself, I’ll stick with all four chips and disable XMP for now. I just want to make sure it doesn’t hurt performance, especially since Ryzen seems to benefit from quicker memory. Should I keep the components in this setup? They appear to function properly, and given their differences, it might improve compatibility as you mentioned.
You're in a tough spot, but this tool can help verify things. If it doesn't recognize the ICS, check the GSKill version for clues or adjust the CL to 19 at 3800 MHz. Try setting the frequency to 3800 and see if it still works—if not, dropping to 3200 or 3000 might help. For getting the units up to date, run XMP profile checks; if it doesn't sync properly, you may need to tweak timing per channel instead of matching both channels. This setup assumes the Samsung sticks are C-Dies otherwise 3600 should be a safe bet with some adjustments.
The IC code indicates an 8Gb Samsung E die, which makes me uncertain about its performance. It seems more likely to be closer to D die rather than C die, though. Since you're using an APU, it will affect it significantly, especially in CPU benchmarks where memory speed is crucial due to smaller cache sizes.