Cold Boot errors: 2x2GB F3-10666CL7D-4GBECO and a Gigabyte GA-P55A-UD4P

[aced]

After booting from a hard power-down that lasts 8 hours or more, I get thousands of Memtest errors in the first pass.

Motherboard: Gigabyte GA-P55A-UD4P with BIOS ver F6
Memory: G.Skill 2 x 2GB F3-10666CL7D-4GBECO (DDR3-1333, 7-7-7-21, 1.35V)
PSU: Seasonic X-650 (SS-650KM)
Video Card: Gigabyte GV-N96TSL-1GI (nVidia 9600GT)

This is the second set of memory I'm trying for this new build, as the first set appeared to have a bad stick. This new set tests out fine under Memtest86+ v4.00 after 30 passes and multiple warm reboots. However, if I leave the power off for more than 8 hours, I get thousands of errors in the first pass.

BIOS is pretty much at defaults, except that I enabled XMP to Profile1 (only choice) in order to make use of the SPD timing values 7-7-7-21, otherwise the board uses 9-9-9-24.

Do I have faulty DRAM again?

[GSKILL TECH]

Yes, you may need to have the memory replaced if the settings are configured properly.

Thank you
GSKILL SUPPORT

[aced]

What does G.Skill consider as the proper settings for this motherboard and memory?

[GSKILL TECH]

Memory Frequency, timings, and voltage set to the memory's rated specifications.

Thank you
GSKILL SUPPORT

[aced]

Defaults with no XMP enabled show 9-9-9-24 and 1.5V at 1333MHz. The cold-boot problem occurs with these settings.

Enabling XMP sets the memory interface to 7-7-7-21, 1.35V, 1333MHz. The motherboard actually outputs 1.424V, probably because it chose the closest voltage setting of 1.4V. The problem occurs with these settings as well.

I manually set the memory timings and speed to the two different scenarios noted above, in order to bypass SPD and XMP. Still had the cold boot problem.

Memory, motherboard, or something else?

[trekkerxp]

do you have 4 modules of RAM installed or just 2?

I have similar RAM modules and the issues you have posted appear when I have all 8gb installed, but magically disappear when just 4gb (2 modules) are present. By now you have noticed the numerous posts regarding cold boot and Memtest or cold boot and random BSOD.

[Tradesman]

Add some juice to the CPU/MC, also try DRAM voltage at 1.4, ...... and... what's your tRFC at? May need to raise it a bit.

[aced]

trekkerxp, I have 2 x 2GB modules.

Tradesman, please explain what 'juice' is, and why I should raise tRFC to avoid cold-boot problems. I don't know overclocking terminology, and don't see how tRFC relates to the cold-boot faults. Note that the system was running the DRAM at 1.4V when XMP was enabled and I tried this manually as well.

I'm not overclocking, and besides, shouldn't this DRAM function correctly at the base parameters as defined by SPD and the Intel 1156 spec, without getting XMP involved or resorting to manual tweaks?

The purpose of XMP is to allow additional performance or features beyond the base specification, and the purpose of manual settings is to allow for custom performance optimizations, certainly not for reliability on a base configuration with allegedly compatible components.

I guess the question is, does G.Skill require manual parameter adjustments to achieve system compatibility, or is there a faulty component in my system that needs to be replaced?

[GSKILL TECH]

Simply send the memory package in for new replacements. The new set should work better.

Thank you
GSKILL SUPPORT