CML16GX3M4A1600C9B XMP profile fails

Hello, new to the forum here. I believe I've done my homework, but im sure someone will straighten me out if not.

 

Symptoms:

OS installs and boots cleanly when memory is clocked at SPD rates (1333Mhz).

Changing the motherboard's OC tuner to XMP, and choosing the default (and only) xmp profile for this set of ram, the system will not boot, or even POST for that matter.

 

I have read and attempted most of the suggestions made in this thread:

 

http://forum.corsair.com/forums/showthread.php?t=103184&highlight=cml16gx3m4a1600c9b+asus+p8z68-v+gen

including adjusting VCCIO and various memtest combinations.

 

since the specs of the build and the symptoms s/he is experiencing are very similar to mine, differing only in the video card. However, I have also tried using the onboard video and observed no change.

 

My own findings:

Running memtest on each individual stick (only one pass, I'll admit), each stick passes with no errors.

Running all four sticks together produces thousands of errors almost immediately.

 

I tried the "Mem OK" switch on my mobo, as suggested in the other thread, however, the result is that the system boots, but the memory is clocked back to SPD values.

 

I began trying various combinations of sticks in various slots.

Any single stick boots in any slot at SPD values with no problem

Two of the four sticks will boot at XMP values when placed in motherboard memory slots 1 and/or 3 (individually and together as a pair filling both slots).

The other two sticks will sometimes boot at XMP values when in slots 1 and 3, and other times they will not, indicated by a BIOS message stating "OC failed"

None of the four sticks will boot at XMP values when placed in motherboard memory slots 0 and/or 2.

 

BSOD's have occurred intermittently, with no pattern I have noticed. Messages in these BSOD's include "PFN_LIST_CORRUPT", "DRIVER_CORRUPTED_MMPOOL" and "MEMORY_MANAGEMENT". None but the last has occurred more than once.

 

It seems like the answer is going to be that the first memory channel or slots 0 and/or 2 are defective, however, im hoping that there is something that im missing, as this is the first time ive personally tried to do any kind of OC'ing, even something as simple as setting XMP for memory.

 

Thank you for any and all help you can offer.

Misery loves company I see.

 

Did you try the DRAM Voltage at 1.55? I thought I had with no success, and wanted to see if you had the same result?

 

This is very baffling. I had mine working for 48 plus hours with no issues. When the issue occurs in the link to my post describes I haven't been abel to get the 4 sticks to run in the system.

 

Wish I had a solution for you problem. You certainly documented everything better than I dd (emotional side came through in my post). But glad to know that everything I tried someone else did with no more success than I had.

 

Cheers!

mmmm try to set manually the settings,

install acpi driver from asus dvd (software/aisuite/probe/drvresources/ chose x64 or x 32 os)

shutdown the system

switch on epu and tpu

(if don't boot clear cmos and mem ok switch)

don't worry if the system reboot after post (is the tpu processor)

go on bios , the tpu switch have configured the best values for ram and cpu, on my case 103mhz bclk (i5 2500) cpu ratio 41, ram 16xx mhz, timings 9 9 9 24. I have manually set up the dram voltage to 1.55 and the system works fine on w7 x64, i think can set up the ddr volt max to 1.65 ( on qvl similar memories from corsair are listed with this voltage).

Sorry for my english but i'm italian

Lets see,

 

Nexxsys - No, I don't think i went as high as 1.55 on the dram voltage, but I did try bumping it up a few ticks to see if maybe it wasn't getting enough to power all four slots. I wasn't sure how high was 'safe' with this set of ram. I also tried bumping VCCIO up a bit, as recommended by various other sources I've stumbled across. In the BIOS, the numbers representing voltage are displayed in white up to a certain point, and then they turn yellow once you pass what i can only assume is the 'safe' point. I stopped just before this point on each voltage (individually), but haven't yet tried both at the same time.

 

morion - Ive installed all the software from the motherboard DVD, including AI Suite, as well as going to the product page and getting the most recent version of each (many only had the release version). I have not tried working with EPU or TPU switches at all. I will try these when I get home today. Thanks for the suggestions!

 

The computer has been running since before the initial post (thread was created on it, actually) with 2x4GB sticks of ram in motherboard slots 1 and 3 installed at 1600 speed and 1.5v and slots 0 and 2 empty. No problems in this setup, however I haven't done anything more complex than web browsing on it.

 

I'm hoping to get this figured out before I leave RMA time frame for all of the parts. Clearly the memory would be an easier fix, but its looking like the motherboard is to blame. I have some -much- older ram that I could put in the slots, but I'm not sure how good a test it would be, since the new ram works fine under SPD rating of 1333Mhz, with all four slots installed, and the older ram does not have any kind of XMP compatibility, so its not a particularly close approximation.

In have already performed your same tests..... 3 slot of ram all ok!!! 4 slot bsod ...... change all ram bsod ..... invert ram all ok ..... test single bank with mem test and all ok |2 slot ok 3 slot ok |put 4 slot and some errors and crashes........ the acpi driver don't install on automatic when install aisuite..... install that manually, the epu,tpu and 1.55 volt is already a try but have worked.....imho 1.65 is safe...... over not i think

Ok, I probably do not have the ACPI driver installed then. I will add that to the list of things to try this evening.

Ok, I probably do not have the ACPI driver installed then. I will add that to the list of things to try this evening.

 

Sorry but why would the ACPI driver (a windows item) be required to have the system boot with 16GB? That doesn't make sense to me. :confused:

 

Also how would this change that 1 oout of the 4 is the problem?

Sorry but why would the ACPI driver (a windows item) be required to have the system boot with 16GB? That doesn't make sense to me. :confused:

 

Also how would this change that 1 oout of the 4 is the problem?

 

same thinks..... but i have debug the bsod errors ( memory management ecc) and this type of bsod ( on my case ) is caused by the acpi driver of windows..... if install the asus specific acpi driver the system work with 16gb ddr3 1600....

if you can post the bsod file i can debug it

So I have tried the approach that morion suggested above.

 

I browsed the contents of my mobo driver cd, and found the file he was talking about. It is not specifically listed in the menu that is displayed on autorun, it is merely listed in the files on the DVD. I ran the 'installer', it appeared to pop up a shell and execute, then close immediately. I then shut down, put the memory sticks in slots 0 and 2, as these were the two that previously would not boot when populated and XMP enabled. I also enabled both the EPU ant TPU switches on my motherboard, then hooked it all back up and booted. 3 system initiated restarts later, the system boots.

 

All seems well for a few minutes, i load up CPU-z to see if the ram is indeed set to 1600, and i THINK it is. DRAM frequency reads 816MHz, which, if I remember correctly is higher than the 667 that memtest displayed when i was at 1333 SPD rate. (forgive me for not knowing the difference between the two sets of numbers, 1600/816Mhz, 1333/667Mhz). The values for cpu ration, BCLK, ram speed and timings have all been auto-set to new values.

 

I shut down, then manually turn it back on to check the bios, but the system goes into the same routine it did when I first enabled EPU and TPU switches, that is, it started up and shut itself down several times before finally POSTing and booting (cleanly). I shut down and did this one more time, finally catching and going into bios before it hits Windows, and the settings seem to reflect XMP values.

The third time booting it came to windows cleanly.

 

Next, I tried installing the other two sticks of ram into slots 1 and 3.

First boot with 4 sticks of ram, with TPU and EPU switches enabled, the MemOK LED lights up. I push it and hold for 3 seconds, initiating the procedure that finds the optimal setup. After completing the process, the system is been clocked back to 1359 rate.

At this point I decide to try manually setting various rates:

1632, voltage up to 1.55, VCCIO up to 1.1, No boot, system resets values to 1359.

1632, voltage up to 1.5625, VCCIO at 1.1, No boot. system resets values again

1648, voltage up to 1.575, VCCIO at 1.1, no boot. system requires mem ok button intervention to boot. values reset eventually.

1672, voltage at 1.55, VCCIO at 1.1125, no values reset.

1648, voltage up to 1.55625, VCCIO up to 1.11875. No boot, mem ok switch required..

 

I chose my values from the drop down list in the bios, and each time the system reset values to something 'safe', the list of options changed by a few Mhz.

 

Memory sticks are warm to touch at this point, but by no means hot. Of course, ive done nothing more than boot and reboot, but thought it worth mentioning.

 

So I think ive addressed all the suggestions so far, but im open to further ideas.

Wow. Thanks I wish there was success. Here is the last thing I have found that I was going to try, but because 1 of my 4 is showing errors I am skeptical. Get the system to Post and set the settings in Bios to XMP but bump the DRAM voltage up to 1.64v. Sounds strange more than few on Newegg and similar buying sites have had success with our boards and that voltage when using 4X4GB sticks.

 

Was the last thing I hadn't tried.

So I have tried the approach that morion suggested above.

 

I browsed the contents of my mobo driver cd, and found the file he was talking about. It is not specifically listed in the menu that is displayed on autorun, it is merely listed in the files on the DVD. I ran the 'installer', it appeared to pop up a shell and execute, then close immediately. I then shut down, put the memory sticks in slots 0 and 2, as these were the two that previously would not boot when populated and XMP enabled. I also enabled both the EPU ant TPU switches on my motherboard, then hooked it all back up and booted. 3 system initiated restarts later, the system boots.

 

All seems well for a few minutes, i load up CPU-z to see if the ram is indeed set to 1600, and i THINK it is. DRAM frequency reads 816MHz, which, if I remember correctly is higher than the 667 that memtest displayed when i was at 1333 SPD rate. (forgive me for not knowing the difference between the two sets of numbers, 1600/816Mhz, 1333/667Mhz). The values for cpu ration, BCLK, ram speed and timings have all been auto-set to new values.

 

I shut down, then manually turn it back on to check the bios, but the system goes into the same routine it did when I first enabled EPU and TPU switches, that is, it started up and shut itself down several times before finally POSTing and booting (cleanly). I shut down and did this one more time, finally catching and going into bios before it hits Windows, and the settings seem to reflect XMP values.

The third time booting it came to windows cleanly.

 

Next, I tried installing the other two sticks of ram into slots 1 and 3.

First boot with 4 sticks of ram, with TPU and EPU switches enabled, the MemOK LED lights up. I push it and hold for 3 seconds, initiating the procedure that finds the optimal setup. After completing the process, the system is been clocked back to 1359 rate.

At this point I decide to try manually setting various rates:

1632, voltage up to 1.55, VCCIO up to 1.1, No boot, system resets values to 1359.

1632, voltage up to 1.5625, VCCIO at 1.1, No boot. system resets values again

1648, voltage up to 1.575, VCCIO at 1.1, no boot. system requires mem ok button intervention to boot. values reset eventually.

1672, voltage at 1.55, VCCIO at 1.1125, no values reset.

1648, voltage up to 1.55625, VCCIO up to 1.11875. No boot, mem ok switch required..

 

I chose my values from the drop down list in the bios, and each time the system reset values to something 'safe', the list of options changed by a few Mhz.

 

Memory sticks are warm to touch at this point, but by no means hot. Of course, ive done nothing more than boot and reboot, but thought it worth mentioning.

 

So I think ive addressed all the suggestions so far, but im open to further ideas.

 

When you enable tpu and epu the mb during post try to set the best values on blck dram freq ecc... reboots are normal, the values of the boot are the same on my system and its work perfect, i have 103 blck 41 ratio 9 9 9 24 timing and 1600 or above dram freq same values of xmp profile, the only change i made is 1.55 to dram voltage.

When you enable tpu and epu the mb during post try to set the best values on blck dram freq ecc... reboots are normal, the values of the boot are the same on my system and its work perfect, i have 103 blck 41 ratio 9 9 9 24 timing and 1600 or above dram freq same values of xmp profile, the only change i made is 1.55 to dram voltage.

 

I understand. I tried exactly that, but it did not work. After i tried that, i also tried the various settings that i listed.

 

After posting last night, i also tried a few other things. I tried booting single sticks in slot 0, with EPU/TPU enabled. Two of the four sticks booted into windows with dram freq of 1600, as identified by CPU-Z program. The other two would only boot after several cycles of the auto tuning process. CPU-Z identified a dram freq of 1333 for those two sticks.

 

So, four single sticks, two boot at XMP values, two do not. All four boot at SPD values. I then took the two that worked under XMP and placed them in slots 0 and 1 at the same time, and the system would not boot under XMP values. Had to use the MemOK button to boot at all, at which point the values were back at SPD.

 

I will try what Nexxsys suggested next. Thanks for your continued help.

i think pump up dram voltage to 1.65 is last chance..... this value is the limit of the cpu..... a cmos clear and test again with all ram stick ad tpu and epu on is a possibility if system boot go into bios and set dram voltage step by step (making tests with memtest and inte burn test) to find optimal setup ....

Whats the best method to clear the CMOS on this mobo? Just pull the battery?

 

I have not yet been able to POST with all 4 sticks intsalled and EPU/TPU enabled. Each time i try, it goes through several iterations, and ultimately settles on increased BCLK, CPU multiplier, but stock ram voltage and dram frequency. Only THEN does it post.

 

Im going to try memtesting 3-4 passes of individual sticks, with EPU/TPU enabled, and see if anything crops up.

 

I will also try memtesting 3-4 passes of individual sticks with EPU/TPU disabled, but OC profile set to XMP, so that there is no CPU overclocking done.

 

I have already done single passes of memtest at stock values, and found no errors.

Whats the best method to clear the CMOS on this mobo? Just pull the battery?

 

I have not yet been able to POST with all 4 sticks intsalled and EPU/TPU enabled. Each time i try, it goes through several iterations, and ultimately settles on increased BCLK, CPU multiplier, but stock ram voltage and dram frequency. Only THEN does it post.

 

Im going to try memtesting 3-4 passes of individual sticks, with EPU/TPU enabled, and see if anything crops up.

 

I will also try memtesting 3-4 passes of individual sticks with EPU/TPU disabled, but OC profile set to XMP, so that there is no CPU overclocking done.

 

I have already done single passes of memtest at stock values, and found no errors.

iterations of these settings is normal when tpu is enabled, this processor automatically detect the bes and stables oc values for ram and cpu, the system reboots some times during post.

To clear cmos on the back pannel of mb is located a switch, you can push this microswitch or disconnect the power cord and remove the cmos battery.

i have done same tests and on single sticks no error 2 sticks no error 3 sticks no error, if put all sticks some errors ..... only way for boot with all 4 sticks and 1600 mhz is tpu + epui enabled and 1.5.5 voltage on ram, i have also install the acpi driver of asus dvd for resolve some bsod issues of memory management caused by the system sleep and hardware sleep.....

 

sorry for bad english but i'm italian....

iterations of these settings is normal when tpu is enabled, this processor automatically detect the bes and stables oc values for ram and cpu, the system reboots some times during post.

To clear cmos on the back pannel of mb is located a switch, you can push this microswitch or disconnect the power cord and remove the cmos battery.

i have done same tests and on single sticks no error 2 sticks no error 3 sticks no error, if put all sticks some errors ..... only way for boot with all 4 sticks and 1600 mhz is tpu + epui enabled and 1.5.5 voltage on ram, i have also install the acpi driver of asus dvd for resolve some bsod issues of memory management caused by the system sleep and hardware sleep.....

 

sorry for bad english but i'm italian....

 

It is quite alright, I understand you. But the fact remains that I have done exactly as you have suggested, and it does not work for me. I do appreciate the help, but I'm hoping someone comes along with an official word or alternative suggestion.

mmmm and boot with 2 sticks and set up bios settings manually shutdown and put rest of sticks?

Ill try it tonight.

It is quite alright, I understand you. But the fact remains that I have done exactly as you have suggested, and it does not work for me. I do appreciate the help, but I'm hoping someone comes along with an official word or alternative suggestion.

I think this is an issue with the MB and not the memory. But thats my .02$

Any time you push the mem-OK button on an ASUS P67/z67 Motherboard, the Memory speed is going to reset to 1333Mhz. That is the default speed of the Sandy Bridge CPU's memory controller. Anything above 1333 is considered an overclock to the memory.

 

The Sandy Bridge (2nd gen Core processors) memory controller is on the CPU. Running all DIMM slots full (4 DIMMs) places an additional load on the memory controller. Having your Bclk bumped to 103 does not help memory stability. The memory controller is designed to, and runs at the same base clock as the CPU.

 

The ASUS boards don't always reset the Bclk to 100 (during a CMOS reset), but 100Mhz is what the CPU's designed default is.

 

To give you an idea of how memory speed is related to Bclk (using 1600Mhz memory):

 

When you have a Bclk setting of 100, the motherboard uses an 8x multiplyer (8x100)=800Mhz.. x2(DDR; Double Data Rate) = 1600Mhz effective memory speed. With a 103 Bclk, you are asking the memory to run at (8x103)=824 x2 (DDR) = 1648 Mhz. This can usually be compensated for by the memory controller, with only 2 DIMMs, but rarely with 4on the SB platform. Under this senario, you have asked the IMC (Integrated Memory Controller) to run faster than it is designed for (100Mhz; which it can), and manage 4 DIMMs (which it can also)... but probably not at the same time.

 

Start over resetting everything to default. then go into your UEFI and set Bclk to 100, memory to 1600, and try to boot. If that fails, go into DRAM timing, and manually set the timings for your DIMMs to the 9,9,9,24 2T. If that fails, bump DRAM voltage to 1.53V.

 

You are attempting to mess with too many things that are not going to help your memory stability.

 

Also, make sure you have the latest BIOS for your motherboard. Some early ASUS P67/z68 BIOSs had issues with memory stability.

 

I have Corsair 1866 memory and can not select the XMP profile for it, my system won't boot. But manually setting everything works like a charm. I am even able to run them at slightly tighter timings than they are designed to run. But I am only running two DIMMs.

It seemed that way from the get go, but i was hoping someone could give me a foolproof test to be sure.

 

Anyway, I was going to preform all sorts of tests tonight, but I can save myself the hassle.

 

Bent motherboard pin.

 

sigh.

 

off to deal with the manufacturer.

 

Thanks for all your help all.

I recently bought a similar system to build (Asus P8Z68-V Pro Gen 3 like the link you referenced), but with the Corsair CMZ8GX3M1A1600C10 (Vengeance 8gig 1600) and I was having the exact issue: BSOD and random memory read errors, I could get 2 sticks to boot but 4 got wacky, etc.

 

I went into the bios and set the XMP profile, but made one alteration that seems counter intuitive: I lowered the RAM voltage to 1.475 volts. Voila! My system was instantly stable with both memory tests and RAM hungry programs like Adobe After Effects.

Interesting. What made you think to drop the voltage?

Interesting. What made you think to drop the voltage?

 

I tried some other things including dropping the clockrate and the BCLK frequency and nothing gave me a different response one way or the other.

 

Originally, my RAM was set at 1.5 volts but the "actual" volts read 1.515 so my system is adding a little overhead voltage already, no idea why.

 

I saw the reply on this thread to try 1.55 volts, nothing. I bumped it up to about 1.575 volts and received more memory errors -- great, at least something has changed.

 

Dropped the voltage down to 1.475, things seemed stable. Ran some memory diagnostics, no errors. I rendered some video in After Effects as a "real world" test (my computer is primarily a video workstation) and things held steady.

 

Long story short, just a matter of elimination.