BSODs with XMP activated, assistance needed.
BSODs with XMP activated, assistance needed.
Hello everyone! I’m facing a problem with my new PC and haven’t found any solutions online, so I joined the forums to see if others could assist. I built my first PC in nine years to replace an old one. The setup included these components: an i5-13600KF processor paired with an MSI Z790 MAG Tomahawk DDR4 RAM (wired, 3600MHz), a Corsair Vengeance RGB Pro 4x16GB Wi-Fi stick at 3200MHz, and the Radeon 6800XT graphics card. If you need more details, just let me know. After assembling everything, I encountered instability. I thought it was due to an old Windows installation, so I reinstalled Windows. That fixed the issue—Windows stopped crashing with BSODs.
I also disabled XMP for the RAM and ran a memory test. With XMP off, everything worked smoothly at 3200MHz. My BIOS was mostly stock, but I changed the CPU load setting from 12 to 6 to reduce overheating.
My question: the i5-13600KF claims support up to 3200MHz. Can I still get it to run at my original speed? Or did I miss something in the process? Thanks for your time and advice!
Certain boards encounter issues with the xmp profile. Adjusting the timings can help determine if the problem is resolved. Are you using the most recent BIOS version?
Check for the latest version of the bios. If XMP isn't working, try running them manually at 3600.
Initially, testing a couple BIOS updates can be a smart move, especially if memory support has been compromised before. If you haven’t already, consider applying some BIOS upgrades. When preparing for configuration changes, keep an eye on these aspects. First, identify the memory ICs in each stick. You likely purchased two 2x16GB kits instead of a single 4x16GB set. With separate kits, mismatched ICs are common, particularly at higher speeds, and this can lead to compatibility problems. The simplest check is removing all DIMMs and checking their version numbers on the labels—something like "ver. 4.31" or "ver. 3.41" indicates the RAM type. If all sticks match, it’s probably not a major concern, though rare issues still exist. If they differ, that’s likely the root cause. Regardless, sharing the version number helps because different ICs may require slight adjustments (like voltage changes or timing tweaks). If the ICs are identical, examine how the sticks are arranged. Dual-rank DIMMs often cause trouble, but single-rank can still be problematic if the board misconfigures voltages. For example, a CPU might expect low VDDQ TX, but the board forces high levels, leading to instability. Try enabling XMP in Windows and take a screenshot of "MemTweakIt" (ASUS tool) to view memory timings. This can reveal unusual settings that might be contributing to the problem. Various sources exist for this, but the most straightforward path is through the hwbot forums.
I've updated to the newest BIOS in the hope it would help. It didn't make a difference. I'll attempt to adjust the timings manually and follow up. These are indeed two separate 16 GB double configurations from different suppliers. I'll share the version details as I go. I need to review the second part of your message—it's a bit complex on the technical side for me right now.
I tend to go a bit over the top with details sometimes. I’d just grab the MemTweakIt app from the link you shared (it mentions it’s for the Z790 Apex and should fit any LGA 1700 board) and capture a screenshot while XMP is active. Also note the version numbers of each stick, which should cover everything you need. The program should display all the important timing info so the board doesn’t get overly constrained on any specific tertiary timing.
From the images it appears the IC is version 3.31 across all sticks. Except for a unique barcode, everything looks identical. I also attached a photo of the packaging: for the memtweak ss, without XMP—after several attempts, I finally enabled XMP and Windows booted without BSOD. Thanks for your continued support.
I need to mention that we must include quotes so you receive a notification. This happened by chance since I was still on the page. The details are clear: the version number should match 8Gb Micron Rev.B for the IC. This isn't inherently problematic—it usually offers good compatibility—but the main issue is its high density, which means dual-rank modules are required. As a result, you'll likely be interacting with the Intel memory controller to determine the maximum frequency. The positive side is that you should have the ability to adjust certain controller voltages, and hopefully, this will resolve the problem.
Initially, I recommend running a stress test on the memory controller. MemTest86 works well for general memory testing but isn't sensitive enough for controller-specific issues. Instead, use y-cruncher with its VST stress tool, which is designed to push the controller hard and reveal instability. You can launch it via its CMD interface or create a batch file for repeated runs.
Once started, navigate to the BIOS settings and experiment with key voltages such as VCCSA and CPU VDDQ. These should be adjusted under the overclocking section. Begin with VCCSA in "Override" mode for stability, then tweak values gradually. Both VCCSA and CPU VDDQ usually perform best around 1.42V and 1.35V respectively. Start at 1.3V each and increment or decrease to see what yields the longest stable operation.
If you're eager to test one setting at a time, a 1.42V/1.35V baseline is a reasonable starting point. However, keep in mind that recent generations often behave differently with DDR5, so caution is advised. Aim for VST to run for about an hour under stress before considering success. If it fails, consider raising the voltage slightly to 1.4V, though this should be a last resort.
In summary, prepare a memory controller stress test, then fine-tune the relevant voltages until you achieve reliable performance.