Maximus impact - xmp ram weird problem

[dragosmax]

Hi everyone, this is my first post on the ROG forums, so please be gentle .
2 days ago i finished building this:

Asus Maximus Impact
Intel i7 4770K with Coolermaster Seidon 240M
Seasonic Bronze 620W
Asus GTX770 DCUII OC (problems manifest with or without it)
ADATA XPG SX900 256GB SSD
2xWD Green 3TB
Bitfenix Phenom Black
+
2x4GB Kingston Beast (KHX24C11T3K2/8X)
2400mhz (11-13-13-30) as XMP1 and 2133mhz as XMP2

Problem is that the board just won't boot with XMP1 profile aplied, or any other 2400mhz manual setting.
Here is the short story:
- assembled the system, booted for the first time and imediately went into BIOS to only apply XMP1 profile. Applied it...restart..board just keeps restarting on and on. I can see some random Q codes, then 40..03 then restart.
- loaded XMP2 2133, booted fine, installed Win7 x64 and drivers.
- reseted CMOS. Plugged in only one stick, aplied XMP1, then boot without any problems into Windows.
- replaced with other stick....no boot, just restarts.
- so i figured that one stick is faulty. Went back to reseller and exchanged it for a brand new kit, then disapointment...same thing happends! Updated BIOS to latest v1002 but no luck!

Conclusion:
- XMP mode: the sticks won't run at 2400mhz XMP1 together, but they do run on XMP2 2133mhz together. Only one stick out of 2 runs at 2400mhz.
- manual mode: the sticks won't run together at 2400mhz 11-13-13-30-2N even if voltage increased to 1.67V or decreased. Loosened the timings to 12-14-14-32-2N...no luck.
- here is the most WEIRD part, that rules out the IMC: they RUN at 2600mhz with 12-14-14-32-2N and 1.65v.
- another WEIRD part, MB related: i always get about +0.011 of DRAM voltage over the set voltage. example: i set 1.64v but 1.651 is actually aplied. When XMP aplies 1.65v, real voltage after reboot is about 1.661 or 1.662. Is this normal?

So what could be wrong with this MB that runs the RAM at 2133 (XMP) and 2600 (manual) but not 2400mhz (XMP1)?
This happend with 2 new kits of 2.

Thank you in advance.

[Chino]

Welcome to the ROG forums, dragosmax.

Have you tried manually configuring the timings and voltages to run them at 2400MHz rather than use the XMP1 profile?

[dragosmax]

Hi Chino, yes i did that, as mentioned above. Sorry for the long post.
They simply do not want to run at 2400 even with loosened timings, not to mention the ones in XMP, even manually set.
Thing is, both kits i tried run (with some hickups) on my friends ASRock. Not on my Impact.
The Impact only runs them at 2133 or 2600, wich is weird.

[HiVizMan]

Your CPU does not like 2400MHz simple as that. I have seen it many times that a CPU has some black hole in memory frequency. I would run 2600MHz in that case.

[dragosmax]

Thank you for your answer.
I was afraid that the CPU might be the problem. Now that's just my ***** luck. Can't exchange it, can't RMA it for this reason here in my country.
For anyone that might be interested, CPU batch is L316B289.

Anyways, i ordered G.Skill Ares Red 16Gb (2x8Gb) 1600Mhz 9-9-9-24, because Kingston 8gb was a temporary solution anyway. I will be returning the "Beasts" next week, because i don't like things not running as advertised on the box.

What i can't explain is why one stick runs XMP1 2400, and the other does not (both tested in A slot). This happened with 2 kits. Does this CPU black hole manifests itself only sometimes? Or does it manifest only in dual channel mode?

Can someone please confirm that the following behavior is normal?: i always get about +0.012 of DRAM voltage over the set voltage. example: i set 1.64v but 1.651-1.653 is actually aplied. When XMP aplies 1.65v, real voltage after reboot is about 1.661-1.663v. The Bios states that the DRAM voltage icrement is 0.005v, but the above scenarios show a very imprecise voltage aplied.

Thank you

[trihy]

Your cpu can handle 2400mhz just fine.


What cannot handle 2400mhz... it´s the asus mb bios.


It´s funny, I´d been using my 2400mhz sticks on a gigabyte sniper 5 for months, changed to a gene VI, if I set xmp to 2400mhz, system restarts itself every 2 seconds never reaching the boot logo in a mad loop that will kill your hard drive if you dont stop that loop.

So... 2400mhz oc to 2600mhz... with a little more timmings... and boot just fine.

And I though gigabyte z87 bios were a disaster... well.. they are not the only one.

[dragosmax]

Trihy, that's exactly what my rig is experiencing at 2400. Mad loop restarts. And to think that i was sorry for not buying the Gene . Anyway, by the time a new BIOS will be released (if any), i won't have any 2400 sticks to test it with. The 1600 G.Skills Ares are on their way. I hope to squeeze some lower latencies out of them, or at least some 1866 at 1.5v
For my particular mITX scenario, i'm not upset with the board, and would buy it again, as it's by far the sexiest board out there.

So i guess this topic can me marked as solved.
Thank you

[trihy]

Good to know that is just another stupid bug.


Asus motherboards looks great. The gene vi it's amazing. Every single output you could wish to have.. is there. Quality.. top notch..


But... on z87 boards, most companies wasted time making their bios looks cool... and.. hey... they look cool, but the amount of bugs is disgusting. I used to hate gigabyte, their z87 boards were a nightmare from the bios perspective. But since I'd moved to asus... Im experiencing lots of problems too.

Hope they start fixing bugs. Lot of them are gross.

[HiVizMan]

Cheers guys comments are appreciated and welcomed will see what can be done to sort things out for future bios versions.

[trihy]

NIce to see that