HCI MemTest or Windows Memory Diagnostic for RAM Testing (Beyond Memtest 86)
HCI MemTest or Windows Memory Diagnostic for RAM Testing (Beyond Memtest 86)
I've consistently used HCI Memtest for different purposes, as it's easy to launch 16 tests and see the results. I prefer Windows testing since it feels more versatile. Compared to other tools, people often favor HCI Memtest for checking RAM and memory performance. I recently completed two cycles, starting at 409.6mb per test and progressing to 1638.4mb. I also ran a 163.125mb test simulating 26.1gb of available memory. Each run showed no errors, so it seems worthwhile to use the Windows Memory Diagnostic or I can rely on HCI Memtest.
You gather feedback from ten individuals, each giving a unique perspective. From what I've observed, Memory Diagnostics typically identifies problems only after severe issues appear, which is why it isn't part of the standard memory stress tests I conduct. It doesn’t mean you should avoid using it, but it isn’t something I routinely perform. The only reliable method I rely on is conducting several stress tests over extended periods—usually multiple hours. For my current setup with a Ryzen 7 3800 CL14 overclock, I stress-test the main system to 700% HCI MemTest, run TestMem5 Extreme1 for three hours, and perform Linpack with GPU stress tests for two hours. I do this because I’ve seen it work, and if it passes without errors, I’m confident. If it fails during a quick test, I stop there. Running it with DDR5 overclocking would have given me more time—last time with DDR5, I completed 24 hours of testing and still saw failures within ten minutes. So, unless you’re overclocking RAM specifically, sticking with HCI or TestMem5 for 1–2 hours is usually sufficient. If you suspect instability, I’d just follow your existing plan and expand the duration or mix different tests to catch any issues early.
system appears stable with no instability issues. I didn't perform any overclocking beyond XMP/DOCP settings. Just wanted to verify my RAM isn't having hidden problems. HCI Memtest ran for roughly an hour testing 16 instances of 1638.4mb each at full capacity, consuming about 90% of available memory with no errors detected. Other tests using different sizes (163.125mb, 409.1mb, especially 1.631mb) used minimal memory, suggesting the larger test was appropriate. Typically I avoid such forum discussions, but finding a direct comparison online is difficult. Whenever I search, Memtest86 consistently appears. This is generally regarded as the top RAM testing tool, though HCI offers more convenience and is considered reliable despite its slower performance due to needing multiple instances.
I've faced memory challenges before yet, but quickly slip into other situations. It works well though—it's not perfect. To confirm your new RAM functions properly right out of the box, HCI is sufficient. For thorough validation, I'd run more than just HCI for several hours and perform various stress tests. However, it usually holds up. There are alternatives that could simplify this process as well. One option is the DRAM Calculator for Ryzen from TechPowerUp. Although the interface feels a bit outdated and packed with features that aren't essential, it provides a much improved GUI for free HCI use (the paid version is superior). Just navigate to the "MEMBENCH" tab, switch to stress mode, adjust the total RAM you wish to test (e.g., 16GB → 14GB), set your desired threads, choose the duration, and start. TestMem5 seems to be a more suitable choice. The interface is somewhat awkward at first, but it becomes quicker to set up and operate once you adapt.
This RAM is just over four months old. No overclocking has been applied except for enabling XMP. I adjusted the BCLK frequency last month, which led to a CMOS reset because I wasn’t sure what I was doing. Timings and voltages didn’t match, so it’s now back to Auto/stock settings. I’ll keep using HCI for now; my first test was in April and the other tests followed in May. Even though I wasn’t using it perfectly at first, I’ve run many HCI tests with various sizes and loads without any errors. Thanks for the advice—I’m still using the free version, just a few extra steps.
I didn’t mean to answer myself; let go of those sausage fingers.
Additionally, what caught my attention was the frequent crashes of the "Search" application in Event Viewer during the HCI Memtest cycle. This error occurred at least four times, and I’m uncertain if it’s linked to the CPU reaching full usage while RAM was at 90%. It happened consistently throughout the hour.
I should verify this. I understand the Windows Event Viewer sometimes displays random errors, which isn't too surprising. However, it's possible you're overloading your RAM and it's causing the page file to appear. Consider lowering the usage slightly—perhaps around 500MB—and observe the results. At 90% usage, that adjustment might not be practical.
Event Viewer serves as a handy yet general utility. I navigated through it to locate any recurring errors from previous HCI sessions, discovering four instances of Bad_Module_Info issues on June 14th and 15th. Two occurred at 5:58pm and 6:16pm, while the other two appeared at 12:16pm and 12:21pm on the 15th. Despite these occurrences, my games remained stable, resources stayed balanced, and no crashes or performance issues arose. I was playing Dark Souls 3 during or near those times, yet the error no longer appears. I’ve also run DDU and performed SFC/scannow, which restored corrupted files, suggesting a possible file corruption incident.