7900x, x670e aorus master, powerful impact during activation of xmp profile
7900x, x670e aorus master, powerful impact during activation of xmp profile
Hello, I just assembled my first PC using the 7900X, X670E Aorus Master, and the RAM you linked. The product page doesn’t show this exact configuration, but it matches the frequency and timings found in 16GB kits, so I thought it would work when I enabled the XMP profile. Sometimes it fails to boot at all without it. When it does boot, stress testing with Prime95 causes immediate crashes—screen freezes, strange colored lines appear, and sometimes a blue screen appears. I don’t have a dedicated GPU yet, so I’m wondering if the built-in one might be the issue. Also, my BIOS was recently updated to improve memory compatibility, but the problem persists. Please help me understand what’s going on.
Execute MemTest86 for a few runs. The memory might be defective at launch, so consider testing it before making significant changes to your configurations.
6000mhz isn't too bad after all, and it shouldn't hurt performance. Still, I'd like the full 6400 I paid for—maybe adjusting the XMP profile could bring it closer if it works out. Even though memory overclocking is tricky and I'm new to it, I got an XMP profile by accident.
Updates may enhance memory compatibility further. Performance gaps are probably only apparent in synthetic tests. Higher speed generally means more issues could arise. A dedicated GPU might fix it, but it may take some time. 6000 mt/s remains solid.
yea MemTest86 so i think i will just stick with 6000 mt/s for now and hopefully it resolves it self with bios updates or when i get my gpu, because it does seem the integrated gpu is failing (because of the weird colors and stuff but maybe im wrong). anyways the performance benefits are probably not worth my time learning to oc memory and spending the time doing it. kinda suck but oh well. any suggestion for stress testing the system for stability? prime 95 for like 8 hours is what i know, good enough?
The 16GB memory listed for Gigabyte supports one stick only. For your model F5-6400J3239G16GA2-TZ5RK, Gigabyte also lists it as F5-6400J3239GX2-TZ5RK. Both G16GA2 and G16GX2 refer to the same configuration: https://www.gskill.com/product/165/374/1...6GA2-TZ5RK.
The XMP 6000 36-36-36-96 setting doesn’t function properly on the same board with the original BIOS (F6) or even with the latest BIOS (F7). It works correctly with the newest BIOS version F8a.
You mentioned the 6400 MHz model isn’t compatible. It’s unclear what your system’s maximum capability is—some sources suggest only 50% run at 3200 MHz and the other half at 3100 MHz (DDR5-6200). Many reviewers didn’t produce 6400 EXPO kits, possibly due to performance concerns.
I can’t assist with troubleshooting crashes since my system isn’t fully stable. My BIOS is slow; keyboard and mouse respond poorly. The RAM stability is also an issue—even at 4800 CL40 I experienced overheating (31 minutes onward). My last experience was a BSOD after 9 minutes in Karhu, with temperatures reaching 63.8°C. How about yours? Are Hynix components cooler than advertised? On the 4800 CL40, I reached 50°C after 23 minutes. Yesterday I had a crash in Karhu after 9 minutes. The problem occurred twice when switching from XMP to 4800 CL40: first boot with BSOD before login, second boot with stuttering sound, third boot with BSOD before login, and fourth boot usually works but still had audio issues. I’m currently using BIOS F7, which caps performance at around 4800 CL40 (possibly up to 5600). I’m testing if the BIOS itself is slow, which would be frustrating. On my Z590 + 10900K + 4000 board, dual booting caused issues with timing changes; Samsung RAM also overheated at 54°C, and I couldn’t lower timings due to heat. Now I realize this was less severe compared to the newer Zen 4 models. A temperature of 64°C is definitely too high. I’d need to replace the RAM for a Hynix 6000 model (30-38-38) or check which type performs better. The EXPO and XMP profiles shouldn’t matter much since they share similar timings; tWR 90 and tRFC 884 are more relevant. Another 300€ would be a waste if the fix is simple.
They offer fresh units featuring AMD EXPO OC profiles such as TridentZ5 Neo and FlareX.
EXPO isn't the key here; this idea needs to be challenged now. It seemed quite amusing before the Zen 4 launch when many believed special divine settings were built in. My 6000 36-36-36 XMP configuration: Frequency - 6000 tCL - 36 tRCD/tRP - 36 tRAS - 96 Vdd - 1.35 Vddq - 1.35 Vpp - 1.80 Vimc - 1.35 tRC - 132 tWR - 90 tRFC - 884 tRFC2 - 480 6000 30-38-38 EXPO guidelines: https://www.geeknetic.es/Review/2428/GSK...eview.html Mostly unchanged except for a minor tRC tweak. These are two distinct chips—Samsung 36-36-36 versus Hynix 30-38-38—and yet they share identical timings. It's surprising how few unique settings exist in XMP/EXPO, mainly just tWR and tRFC. The board usually applies 90 and 884 automatically, meaning XMP/EXPO doesn't offer special perks; it's just standard specs. These figures match what you'd find on official RAM pages, stick labels, and product listings. This pattern has persisted since the DDR3 era when XMP was first introduced. I did notice an odd issue though: without enabling XMP, UCLK drops to 1500 MHz even if you set it to Auto or MEMCLK, and it fails at 3000 MHz without XMP active.