F3-12800CL9D-8GBXL advanced timings

[Logicaluser]

intel i3 540

F3-12800CL9D-8GBXL

Just installed an 8GB kit and noticed a few odd things while setting timings in BIOS

I enabled XMP, set my memory multiplier to 5x (160 bclk), timings to auto and memory voltage to 1.5v. Upon reset it shows 9/9/9/24 timings in BIOS, but no detail on what subtimings are in use, just those specified in SPD & XMP. Checking from within windows the following don't seem to match the XMP profile:

tRFC 175
Command Rate 1T

Should I set all the subtimings to match those detailed in the XMP profile?
This BIOS is odd in that on the manual subtimings page there is no "auto" or mention of what 'default' is for given values.

current timing info:


Uploaded with ImageShack.us

[GSKILL TECH]

The timings are fine.

Thank you
GSKILL TECH

[Logicaluser]

Left it running prime95 and came back 20min ago to the computer having reset and waiting at the bootloader's OS select screen. Windows says it was not shut down correctly, however there is no minidump file created

[GSKILL TECH]

OK, then try XMP settings. The settings you had were slightly tight, so if it was stable, then it was OK, but since unstable, try XMP with Command Rate 2T and it should be fine.

Thank you
GSKILL TECH

[Logicaluser]

Board isn't applying XMP timings when set to auto, manually set all timings listed in XMP profile.

Does this look right?



Uploaded with ImageShack.us

[Tradesman]

Looks pretty good, but what you do is change from AUTO to Enable XMP then select the profile

[Logicaluser]

Again I came back to it reset after running prime95, using settings in previous post

I reset the CMOS, and set BIOS from scratch.

Enabling XMP adjusts my bclk to 160, and brings up a 2nd option with two choices, Auto & Advance (appears to lock/unlock memory multiplier). Neither appears to influence subtimings. There is a memory-z option in BIOS that shows me both SPD timings and the XMP profile timings, but I see no particular way to 'apply' the XMP subtimings automatically.

[Tradesman]

Prob going to have to see if GSkill has info on it, should be back Fri

[Logicaluser]

I managed to get the system 2hr prime95 stable last night by setting the DRAM voltage to "Auto" instead of manual 1.5v

In this BIOS it seems rather than "Auto" being an on/off toggle it shows Auto as being a "position" between two voltages ( 1.597v and 1.603v currently) And yet... the description on the right in BIOS shows Auto as 1.513v Edit: Prior to this RAM upgrade the DRAM voltage "Auto" appeared as a position in between 1.498v & 1.503v

So... being confused by this discrepancy I realized MSI offers a windows utility for the mobo, so I grabbed ControlCenter to check what actual DRAM voltage is in use when set to "Auto" and it shows 1.513v when booted set to "Auto". On a hunch, I set DRAM voltage back to manual 1.5v and upon boot ControlCenter reported a DRAM voltage of 1.7v. I've switched back to "Auto", and its booted w 1.513v now.

Has enabling XMP screwed with what my board considers "Auto"? How/why did manually set 1.5v = 1.7v?
Should I just forget about XMP and do it ALL manually?

[Logicaluser]

Thus far I'm prime95 stable for 2hr with the settings in image below

Next step will be to set tWR tWTR tRRD tRAP to match the XMP profile data
however... I don't see a place to adjust tRC in BIOS and the BIOS has tFAW which does not appear listed in XMP/SPD data in the apps I've tried.
tFAW is currently set to 63, but no clue where it got this value tho, MSI control center shows this as the max value, and in the red.

Edit: Google finds me several reviews talking about tFAW in other 2x4GB DDR3 kits, and the values mentioned are WAY below 63. Is there a recommended value for this kit?



Uploaded with ImageShack.us

[GSKILL TECH]

AUTO can change every time you boot up, that's why we always recommend manual settings.

tFAW 24

Thank you
GSKILL TECH

[Logicaluser]

AUTO can change every time you boot up, that's why we always recommend manual settings.

tFAW 24

Thank you
GSKILL TECH

Thanks for the info

I agree that manual settings SHOULD be the way to go, however it seems with this MSI motherboard's BIOS and XMP it's not that simple.

When I specify a manual 1.5v RAM voltage with XMP enabled the result is DIMM voltage of 1.7v read from windows (just prior to crashing)
Why, I've got no clue, but it does. (manual 1.5v did NOT do this when using previous non-XMP RAM)

Meanwhile, DIMM voltage set to "Auto" with XMP enabled gives me 1.513v consistently when read from windows, even tho BIOS itself currently says auto = 1.6v in one place and 1.513v in another

IMO it would be useful to have some sort of warning for others using similar MSI boards to watch out for such behavior when using XMP, for fear of frying some RAM with 1.7v when set manually to 1.5v. At the very least a recommendation to check from windows to make sure DIMM voltage has been properly applied
The H55M-P31 is a MSI-7636 board (H55M-P33, E23, E33 & E55 are variants)


Will set tFAW to 24. System is currently 3hrs Prime95 stable with all other timings set manually to match XMP data, and DIMM voltage to 'auto' (1.513v)

[Tradesman]

I'd check with MSI and explain what's going on, they may have a beta BIOS that might help or be able to tell you if they are working on it. Doesn't appear it's reading the XMP off the sticks correctly or if it is it isn't able to apply the settings. May be mobo itself or BIOS.