Intel i5 6600K random lockup Freeze or BSOD...

[satas]

Hi to all!

firstly here is my configuration:

Intel i5 6600k at stock speeds

Asus Z170 ROG VIII Gene (latest bios 1504)

Corsair DDR4 16GB (2x8GB) 2400MHz (CMK16GX4M2A2400C14)

EVGA GTX980 SC (at stock speed)

I am experiencing random system freezes and BSODs. The screen locks up and the sound is stuck or it is just BSOD.

I already started a thread over at Tenforums and did a lot of troubleshooting but nothing really helped.

most of the time the issue happens while just browsing the internet, playing some flashgame in the browser etc.

it can randomly happen with listening to music with wmp or aimp or just while starting a portable application from usb drive.

it's really very random and unpredictable, that makes it very difficult to analyze - but the issue is there in the background all the time.

I'm running Windows 10 Pro and have the all the latest bios and drivers installed.

there is no manual overclock except the ram running on XMP profile at 2400MHz (instead of 2133Mhz).

how is it possible that Intel's Skylake can not handle ram at 2400MHz while Z170 chipset advertises speeds at >3000MHz.

Thank you in advance for the help and advices,

Sahin

[CCris2]

Hi,

From what I saw on asus website, they have a firmware update for that asmedia controller.

It is not enough to disable Asmedia usb in bios, You need to remove the drivers if there is a problem with them.

I have ddr4 3200 and I have no problem with it. Only problem is the video drivers. Intel support team is working hard I think for those drivers as it seems they have a lot of problems.

So there is no problem with the high speed. Your Ram speed is low compared to mine.

First of all if you have a intel hd 530 igp enabled, uninstall all drivers and then disable it in bios. Then check if you still have same problems.

If problem still persist try change video drivers, try older or newer drivers because from what i saw there are some buggy drivers for some nvidia cards.

Also there is a possibility that a driver from an external device that you install to create problems on your system.

If you can't fix the problem, there may be a problem with the cpu and the way your cooler is mounted on it. There are lots of posts on internet about fragility of new skylake cpus. People managed to damage the cpu because the mounting system of the cooler they mounted on it had no springs and they screwed it to tight.

I hope some of these advices will help you to fix the problem.

All i can do is wishing you good luck in finding it

[satas]

Hi Ronin,

thanks for the fast reply and the suggestions.

prior to disabling the ASMedia controller in the bios I uninstalled the driver under Windows 10.

I suspected the driver as causing the issues but it was not that.

furthermore I have never used the iGPU, it's disabled in the bios and I never installed driver for that.

I always use my PCIE video card from the beginning on, driver version did not have much (if any) impact on the issue.

no matter that I always use the latest, inbstalling only the driver and Physx as a clean install.

I'm using a CM Hyper 212 EVO and I payed attention not to screw it too tight.

I can even rotate the cooler slightly so it doesn't apply too much pressure on the cpu.

on one side there are plenty of guys like you using ram >3000MHz with no issue.

on the other hand there are a lot of user failing even at SPD settings.

and some people like me get troubles as soon as they enable XMP.

but we all have something in common: Skylake cpu and chipset - the rams doesn't really seem to matter much.

after reading some other threads I'm almost sure I have the same issue, just not as severe as some others.

http://www.tomshardware.co.uk/forum/id-2830772/skylake-build-randomly-freezing-crashing/page-7.html New Skylake Build Randomly Freezing/Crashing - Page 7 - Intel i7 - Windows 10

here is my thread at Tenforums regarding the BSODs (ram and driver troubleshooting):

http://www.tenforums.com/bsod-crashes-debugging/40664-bsod-attempted_write_to_readonly_memory-windows-10-pro-64bit.html Windows 10 Forums

here is my thread at Asus ROG forums (ram, bios and voltages troubleshooting):

https://rog.asus.com/forum/showthread.php?83392-Asus-Z170-Gene-strange-memory-issues Asus Z170 Gene strange memory issues...

[CCris2]

Try to increase voltages on PCH+1.0V to 1.110, VCCIO to 1.115, VCCSA to 1.25, and check if the Memory voltage is correct. All those voltages are connected to system stability when the memory is overclocked or with tight timings.

These voltages I think should be enough for DDR4 2600.

[satas]

I've been through all that stuff many times Ronin, I've been troubleshooting for months.

it came out that no driver is causing an issue and my rams are also failure free.

while some actions increased stability a bit there doesn't seem to be a permanent solution.

this leads me to the Skylake C-State problem mentioned in lots of threads.

the symptoms are almost same as mines...

here is my thread at Tenforums regarding the BSODs (ram and driver troubleshooting):

http://www.tenforums.com/bsod-crashes-debugging/40664-bsod-attempted_write_to_readonly_memory-windows-10-pro-64bit.html Windows 10 Forums

here is my thread at Asus ROG forums (ram, bios and voltages troubleshooting):

https://rog.asus.com/forum/showthread.php?83392-Asus-Z170-Gene-strange-memory-issues Asus Z170 Gene strange memory issues...

[FHeik]

Your RAM model isn't listed in ASUS' http://dlcdnet.asus.com/pub/ASUS/mb/LGA1151/MAXIMUS_VIII_GENE/M8G_DRAM_QVL-081815.pdf QVL for your mainboard, so you can't expect it to work flawlessly (if at all).

Try a supported model and check if the problem persists.

[satas]

all the troubleshooting described in the threads above was done with the ram CMK16GX4M4A2400C14 (4x4GB), which is on the Asus QVL list.

after finding no solution and suspecting the ram I switched to the CMK16GX2M4A2400C14 (2x8GB).

although I have less issues now the main problem is not gone, it's just more random than before.

that made me think I was facing at least two issues rather than one, if one was the ram the second one seems to be the Skylake C-States.

at least that's my status today...

[FHeik]

One thing I've observed with several Z170 chipset-based systems is, they can be VERY picky when it comes to RAM.

I've even had a case where a RAM model listed in the board manufacturer's QVL caused a freezing scenario just like the one you (and many others) are facing, although every RAM test I conducted didn't indicate any error, with the sole exception of a few errors in MemTest86+'s test 13 (Rowhammer; < 10 errors over a period of ~ 5 hrs).

In this particular case, switching to a different RAM brand / model fixed it. But to obtain that insight was a painful odyssey of trial & error on both hardware and software level.

[satas]

I agree completely and I even would do it if I owned a ram store myself.

Skylake is definitely very picky regarding ram, that's why I originally bought a ram suggested in both QVL lists, Asus and Corsair.

but even that did not help at all as I see now, my current ram (not in QVL) performans better than the first one.

I tested the ram over 9 hours with zero errors (MemTest86+), run Windows Driver Verifier over 72 hours etc. (both more than once)

it is an issue that is very difficult to troubleshoot for me as it's random and the only way for me to provoke it a bit is using XMP.

even then it may take a week to happen, but one day it will happen for sure. that's really annoying and disturbing to know.

for me stability is much more important than performance, that's why I bought ram just one step over SPD speeds.

[CCris2]

My only problem is with internet explorer. Is the only one that gives me some kind of errors, but I searched over internet and those errors are from graphic driver ... It seems that there is some kind of memory leak in driver and when there are many pictures, videos on a webpage in ie, in a certain scenario it crashes. This happens all times on facebook when scrolling down thru all those pictures and videos, but only in IE.

On other applicatios or games it is strange but is all ok, well with the exceptions of the games that the driver is not optimized for.

I tested with a discrete graphic card, and those errors are gone ...

I will buy a video card as soon as the new generation comes out (hope they are ok and don't get me into more troubles).

Only way that I can get a blue screen is if the voltages are not set correct when using XMP.

[CCris2]

My system is on Asrock Z170 Extreme 7+ and i7 6700k, and Corsair 2x8 GB 3200 CMK16GX4M2B3200C16R which is not even specified on asrock memory support list but is working.

Only problem was with the first bios. When I was setting memory on XMP system was not even booting. That's when I started experimenting with voltages.

On the next bioses they fixed the problem, and with default settings and xmp I didn't have any issues except those I described in the previous post.

At the moment I have set voltages manually because I saw that the voltages that are set by the mainboard do not correspond with the ones I see in HWinfo ...

[satas]

actually they all should work as long as they stick to Intel's XMP design rules and specifications.

but what if Intel is the actual trouble cause?!

[EstebanA_C_Intel]

Hello:

I would like to know what is the exact code or BSOD code that you are receiving in this scenario.

I look forward to your reply.

Regards,

Esteban C

[satas]

Hi Esteban,

In my previous posts I already included the link to my thread on Tenforums regarding the BSODs.

I'm regularly updating the thread as soon as there is something to report, there are also test results and log files included.

http://www.tenforums.com/bsod-crashes-debugging/40664-bsod-attempted_write_to_readonly_memory-windows-10-pro-64bit.html Windows 10 Forums

Recently I mostly got the "IRQ_NOT_LESS_OR_EQUAL" and the very last one being of type "MEMORY_MANAGEMENT".

The only thing I didn't try out yet that can think of is to apply higher vcore on the cpu, keeping the frequency on stock level.

I'm not sure if it would help, but something for sure makes my system loose stability at one point and it's not the ram.

It is the second ram kit I have similar issues with so I think ram could be eliminated as a cause.

[CCris2]

From what I see in the last HWinfo u posted (I recommend update it), I think VCCSA is low for DDR4 2400. Also I don't see VCCIO.

And CPU OC PLL is on auto or u decreased voltage ? On my mainboard I didn't even touched that voltage, and I think normally should be 0.950 if I am not wrong.

VCCSA and VCCIO are crucial for memory stability if you use XMP.

VCCPRIM 1.0 I think is PCH +1.0 on my HWinfo. That one should have increase voltage also and I see that you have it on standard voltage...

It seems that either your mainboard is not setting correct voltages for the memory you have, or you have set manually wrong voltages.

Please do a bios reset, and update HWinfo.

Your memory has low timings I think on XMP settings and needs more voltage on those points.

And stop checking your hdd. Your system has stability problems mainly from wrong voltages

[satas]

Hi Ronin,

to make the comparison easier I attached my current HWInfo voltage readings. if you do the same we can better compare the values.

but we must keep in mind the bios settings, I run everything on auto currently - no XMP, no OC, no manual voltages, balanced power management profile.

it was also the case in the HWInfo readings you are referring to, which should be identical to the ones I share now.

the reading were measured under Windows 10 Pro with just a browser running in the background, so almost no load.

I use HWInfo64 v.5.22-2820 and I didn't really understand which update you recommended?! is there newer version I'm not aware of?

that are the voltages set by the motherboard on "auto" mode, using XMP the voltages get higher.

no matter that the motherboard could be applying lower voltages than needed.

but in that case wouldn't I get crashes even more often than now?

I updated my bios to the latest version (1601 on Gene) just 1 or 2 days a go.

I didn't set anything manually since then, that's why I didn't reset it afterwards.

my memory has an XMP profile with 14-16-16-31 and I already tried running it with 1.25V instead of 1.2V on XMP.

curreently the ram is running on SPD with 15-15-15-36 timings on 1.2V, set to auto.

but even on XMP the ram should be fine on 1.2V according to Corsair, so I suspect rather the cpu and not the ram.

issues caused by voltage should happen more often and manly under high load, in my opinion.

but that's not the case, the issue is there but very random and never under high load.

the HDD checking was requested on Tenforums because some storage driver related values were in stack as the last crash happened.

I checked the driver and everything seems ok but still miles away from a real solution.

[CCris2]

First of all a portion in the CPU is in fact a controller that manages memory and PCI-E input/output called PCH. Just a point that is not the CPU power.The voltage that powers that portion is called in your case VCCPRIM_1.0.

That portion power on skylake has also and external voltage called VCCSA.

That portion comunicates with the CPU thru signals that need to have a certain voltage called VCCIO. That you can only see in bios ...

All these three voltages are connected.

When your memory has lower timings or higher frequency it needs more power and also needs higher VCCIO voltage because it forces PCH to work at higher frequency. Because PCH works at higher frequency in need also higher power, in this case higher voltage on VCCPRIM_1.0.

VCCSA gives power to VCCIO and I don't know if for PCH also but that means that VCCSA needs to be higher also.

It seems that in your case the mainboard is not giving the right voltages and when your VGA or CPU is doing some operations/data transfers when it need all the power to work correctly the voltage is not enough to assure that power.

So all you can do is either manually set higher voltages, or manually increase memory timings to the ones u see in cpu-z at lower frequencyes.

If you want to increase voltages in your case try increase VCCSA to 1.110V first and see if system is stable. Then VCCPRIM_1.0 to 1.050V, and then VCCIO

to 1.080V.

After u set these voltage if you still get blue screen post here what is the message you get.

PS: When I put memory on xmp for ddr4 3200 my mainboard automatically sets VCCSA to 1.250V.

[satas]

so basically you mean that the memory controller on my cpu is not getting enough voltage and thus leading to crashes.

if this is the case then my motherboard is the issue as it doesn't supply the voltages needed for clear operation on stock speeds.

I could understand higher voltages for OC or XMP but why schould the system need higher voltage on stock speed?

by the way, VCCIO is shown as VTT in HWInfo. when I switch to XMP my voltages also get higher automatically.

what are you voltages on stock speeds, are they higher than mine?

that's how my voltages are when using XMP at 2400MHz.

[CCris2]

Memory is not at stock speed, when is in XMP mode.

That VCCIO is to high for that memory frequency, and I was right

VCCIO is to high and VCCSA to low for that VCCIO.

Try the voltages I posed earlier, and increase also the VCCPRIM_1.0 (from there you get IRQ_NO_LESS_OR_EQUAL error)

[satas]

I'm aware of that, currently I use the ram on stock speeds SPD at 2133MHz.

the last HWInfo I posted is from the last time as I had it on XMP at 2400MHz.

it's the same ram and same rig, the only difference is SPD vs. XMP - just to see the voltage differences.

[CCris2]

I have other CPU and memory so my voltages are higher only that VCCIO is lower rofl