32GB G.Skill F3-1866C10Q-32GSR is not compatible with Asus Maximus VI Hero

[static1]

Hi, the G.Skill site lists G.Skill F3-1866C10Q-32GSR as baing compatible with Maximus VI hero.

This is not the case.

SPECS:
Asus Maximus VI Hero __BIOS.v0903__SouthBridge.v05/C2
32GB (G.Skill F3-1866C10Q-32GSR) 1866HMz (PC-14900) CL10-11-10.30 1.5V (4 sticks x 8GB kit)
Intel i7-4770 (non K version)
Antec TP-550 (fully haswell compatible)

VITALS:
Total power usage is about 50W.
CPU temp 30deg C
Mobo temp 28 deg C


USAGE:
-I am only using cpu, mobo, 32GB ram, and onboard video. (no boot device attached.. no hdd)
-Also using usb keyboard and mice.
-Mobo is on the cardboard box that it came with.
-I am not trying to overclock... just trying to get good stable system before placing new parts in chasis!

WHAT I HAVE DONE:
-I have updated the BIOS. Went fine. Did a battery out and pin BIOS reset. Reset all to defaults (AUTO).
-8pin and main power cable connected.

PROBLEM:
-I am using the red start button on the mobo to start the mobo since it is not in a case.
-It boots fine.

-I see my ram is running at 1333Mhz... so I set XMP to profile 1.
The mobo at this point detects the ram as 1866HMz 10-11-10.30 1.5V - 1.25V
(This seems mostly right except I don't know what the 1.25V is for...??? this is 1.5V ram....)
Is this an issue?

-Then I choose save and reboot option.
The funny thing is it reboots into EUFI bios very fast. All seems OK.
But at this point there is not way to exit or shut down system. Exit option in BIOS just kicks me back into bios!
So I press the red start button. I only hold it for 1 sec. This shuts down the system.

-Then I press red start button again. This is where starnge stuff happens.
The mobo struggles to start. It enters what i call a boot loop. It tries to boot but after 2 secods or so it shuts down. Then tries again and after 2 seconds shuts down. This cylce repeats about 5 times. Then it sucessfully boots into BIOS. I see that RAM settings in BIOS did not get clocked down from XMP. What I don't get is if it can boot after 5 tries, why doesn't it boot after the 1st try? Why does it allways need like 5 tries?

Interesting observation is that in the boot loop, when it fails, if i look at Q led next to power plug... it always gets no further than DRAM. After a few tries it does get past it and then I have a good boot.

So I thought it was the dram. So here are a few observations about that:
-The DRAM works 100% in an AMD FXA mobo.
-When it does BOOT (it always does boot after about 5 or so tries) I can boot into Memtest and run RAM for many passes at XMP settings with zero errors.
-I tried 1866Mhz at 1.6v ... no difference.
-I tried 1600Mhz at 1.5v ... no difference.
-I tried 1600Mhz at 1.6v ... no difference.
-I tried all of the above frequencies and dram voltages with XMP latancies + 1 to each... no difference.


-I tired AUTO which is 1333Mhz... works perfect (boots always on first try).
-I tired 2 stick of RAM in red slots at XMP spec...works perfect (boots always on first try).
-I tired 2 stick of RAM in black slots at XMP spec...works perfect (boots always on first try).
-Any combo of two sticks in any paired slot works ar XMP spec.

The problem is that 4 sticks only works at 1333Mhz... on a cold boot.
If I let it boot loop 5 or 6 loops it will also boot with 4 sticks at XMP spec and mem tests do not fail.

[supahos]

Suggest nudging VCCIO up a couple of notches. It means you're close if its booting and running eventually. Only other thing to try would be trying single sticks, enableling xmp and seeing if they all pass memtest by themselves. or perform differently.

[static1]

Single sticks and even two sticks work flawless in any slot and any slot combination at stock settings for board and XMP profile 1 for ram. Memtests A-OK.

with 4 dims, only 1333mhz is possible. Latency and dram voltage does not help.

One poster on asus rog forum said that maximus boards do not work with G.Skill ram since last two BIOS versions because there is a cold boot issue. Sounds like what I am experiencing since it is precisely "cold boots" that cause boot looping. Other types of boots, like after setting bios settings and selecting reboot option as well in BIOS do not cause problems with 4 dims. If this is true, G.Skill should remove maximus vi boards from compatibility lists... as losing $350 this way is not a trivial thing.
REF: http://rog.asus.com/forum/showthread...untry=&status=

VCCIO is the io / pcie voltage... no? Is this safe for me to touch? This is a sys for development, i need ultra reliability. This is why I purchased parts that i thought were high-end, paying a x2 price premium for less features. Need 100% stability.

[supahos]

Well you have a couple of options... if it was the last two bios updates that broke it, then you can always get the 3rd newest bios from asus and flash it and see if that fixes things. No way Gskill can be responsible for a company breaking the functionality of their product with a bios update. There are way too many mobos to check for this.

As for VCCIO its basically the memory controller voltage. and sometimes does require a bump when maxing out the memory for the board. I would also run it at least 1.6v for a full board. You haven't lost any money with the right combonation of bios/tweaks it should be able to get stable.

[static1]

Do you think 1.6V won't hurt these sticks? They are rated for 1.5V.

Also, what VCCIO do you recommend?

I don't blame G.Skill... My point was that it ins't compatible. Who, why, since... is another matter.

[supahos]

1.6v won't hurt the sticks at all. you can try stepping it down once we get everything working properly, but I can't see it hurting anything even if you leave it like this for years. on VCCIO I've not played with it on a haswell system enough to know yet (only built one and it was i3 business machine) what the defaults are on different cpus. Just enter whatever the value is when its "stock" raise it maybe .05~.1

[static1]

OK, I'll give that a shot and let you know.

[supahos]

Just seems weird to me that Asus would have broken any Gskill memory with a bios update. Seems they usually communicate with eachother pretty well. If it had been any other brand I woudln't have been surprised but usually asus + gskill is an almost guarantee.

[static1]

Yes, I agree.

I went into the BIOS and I don't see VCCIO.

Here is what I see that may be similar.

VTTDDR...........0.750V
DRAM CTRL REF VOLTAGE ON CHA........Auto
DRAM CTRL REF VOLTAGE ON CHB........Auto

Is VTTDDR the same as VCCIO?

[supahos]

Oh heck just did a little dash of reading and that has changed, I suspect its the system agent voltage you need to nudge, but don't mess with it until emmisary, tradesman, or gskill confirm As I said I haven't played as much with haswell and some of the terminology has changed.

[static1]

OK, thanks. I appreciate you trying to help. Interesting idea though. I think you are prob somehow right. Perhaps with 4 sticks there is not enough voltage in some controller/chip.