Memory Stability, BIOS settings used undervolting (is everything okay?)
Memory Stability, BIOS settings used undervolting (is everything okay?)
Hello,
I’m a bit puzzled because it seems my BIOS might have been under-volting the RAM. It would be useful if others could review what I’ve shared and confirm whether I’m handling this correctly.
My setup includes a custom-built system with an ASUS X299 Prime Deluxe motherboard, BIOS version v3201, and a Corsair Vengeance LPX 32 GB DDR4 3000 MHz C15 XMP 2.0 high-performance memory. All memory slots are fully utilized.
I generally prefer standard performance builds but like to keep things at stock settings. Turbo is turned off, and the clock speed remains at the default 3.3Ghz. Apart from disabling turbo, I leave all overclock and AI parameters set to auto. For some periods I’ve experienced occasional blue screen crashes, usually spaced a month apart. Initially, I thought it might be hardware-related, though drivers could also be involved. Mostly NVIDIA crashes occurred, but not consistently.
Recently, there were two consecutive blue screens, prompting me to investigate further. I performed Memtest86 overnight without errors after three passes over eight hours, then restarted into Windows for deeper analysis. I removed, cleaned, and reinserted the graphics card, updated the BIOS to v3201, refreshed chipset drivers on the ASUS site, upgraded the nVidia driver to the latest WHQL version, and purchased HCI Deluxe Memtest. After an overnight run, two errors appeared. I was somewhat skeptical about HCI but also ran a short OCCT license afterward, which confirmed the issue—both memory modules showed 2 errors within eight hours, while the CPU reported no issues in OCCT Linpack.
When examining the memory, I found the part number and checked the default safe specifications (SPD), including voltage and timing details, as well as the tested values. The voltage for CHAB was 1.2V and the actual reading was 1.35V. BIOS settings were all set to AUTO. There are two DRAM channels with voltage controls in BIOS: Volt CHAB (DRAM AB) and Volt CHCD (DRAM CD). Both had adjustment boxes and displayed actual voltages. CHAB was at 1.2V and CHCD at 1.184V, indicating a slight under-voltage for CHCD. I manually adjusted both to 1.2V and 1.215V respectively, which stabilized the readings after reboots.
I then rerun OCCT with CHCD set to 1.215V, achieving stable results without errors. Graphing the voltages showed a minor dip in CHCA around every minute, so I adjusted it to 1.205V. After saving the BIOS changes and restarting OCCT, the voltage remained steady at 1.2V with no errors detected. I plan to continue testing overnight to confirm stability.
I’m still unsure why the BIOS was under-volting CHCD and occasionally CHAB. It seems like a possible optimization attempt, though the SPD value of 1.2V is close to the actual readings. The HCI Design Memtest also showed some inconsistency—errors fluctuated during thread operations. One instance saw errors spike to 72 before dropping to zero.
I appreciate any insights or suggestions you can provide.
Best regards,