Problem detected after upgrade – memory testing shows issues but system remains stable.
Problem detected after upgrade – memory testing shows issues but system remains stable.
Hello! Recently (god when will it stop) i've upgraded my RAM by getting another dual channel kit identiccal to the one I had initially(both from Corsair), then I kept running into app crashes and BSODS( Link - if curious) It was either too many sticks for the memory controller to handle or RAM sticks were different versions and thus they didn't like eachother. So I've decided to buy a brand new dual channel kit, but bigger, instead of having 2 kits, 4 sticks. So I've bought this: ADATA XPG Gammix D35 32GB DDR4 3600MHz CL18 Dual Channel Kit AX4U360016G18I-DTBKD35 (SK Hynix version) XMP Profile is enabled. I've had no app crashes and no BSODS , but (after past experiences) I've decided to test the RAM anyway, and thus, I've found the errors. I've been running with the new kit for 4 days. Tested using both Memtest86+ and Passmark Memtest86, each found very similar errors, but Passmark version seems to have found many errors only on 1 ram stick whereas on Memtest86+ there are errors on both sticks (judging by the failing address GB range)? Test Results: Passmark Memtest86 (didn't know I could save an html report, should be default to always save it on stick ) Test 8 - 2 - [Random number sequence] Test 11 - 11 - [Random number sequence, 64-bit] Test 12 - 2 - [Random number sequence, 128-bit] Memtest86+ Test 8 - 35 - Random number sequence Test 9 - 1 - Modulo 20, random pattern (Test number - No. of Errors - Test name) Errors seem to appear consistently on bits 3' and 4', with only 1 of the bits corrupt per error, on random CPU cores and random memory addresses, only on Random number sequence-type tests. The errors also seem to happen on later "runs" of the program, like after pass 6+, so quite a long time. Any thoughts? Thank you. PC Specs: System age - Hardware: 11 months System age - OS: 5 months (I went through multiple Windows reinstalls recently...) OS * : Windows 10 22H2 x64 bit, Desktop, Retail CPU: AMD Ryzen 5 5600 3.5GHz (NOT the 2020 5600x, it's the 5600 (non X) released in 2022) CPU Cooler: ID-Cooling SE-224-XT V2 Black RAM: ADATA XPG Gammix D35 32GB DDR4 3600MHz CL18 Dual Channel Kit AX4U360016G18I-DTBKD35 (SK Hynix version) GPU: Sapphire Radeon RX 6750 XT MOBO: ASRock B550 Phantom Gaming 4 ATX PSU: Seasonic B12 BC Series, 80 PLUS Bronze, 750W Storage: 1x SSD WD Blue SN570 1TB(System) + 1x SSD Samsung PM981 1TB(Storage)
Only method to truly determine if a stick is faulty is to test each one separately. The software techniques aren't very accurate. Also, ensure XMP is turned off in case it's causing instability (unlikely but possible) and inspect each stick within its memory slot to identify any dead slots or channels. This often points to a temperature problem, such as overheating RAM or CPU. Using a hair dryer on the RAM can help—if errors increase quickly, it suggests part of the issue is thermal.
I believe the slots are okay. I ran several tests with the 2x2 Corsair kits before, checking the RAM in each slot and seeing no issues. I’ll try testing the sticks individually next—maybe Passmark’s software had a better idea. Also, I’m considering the hair dryer method.
After repeated nightly tests, I've confirmed the RAM slots are functioning correctly with no errors during sessions. However, I've observed inconsistencies when XMP is enabled. On the first stick, I consistently encounter 2 bits of error, while the second stick shows occasional errors—sometimes 4, sometimes 2. In two separate tests, the first stick had 2 errors, but one bit was identical across both tests (bit #51), with the other differing. What should I do next?