Need real consulting here as I don't understand what is broken in my system!!!

[PIRATA!]

Ok, after having got a lot of help here from you guys for my laptop, now I need help for my desktop pc. It is a long post, but I'm having strange behaviors here that I don't know what to think about it.

I have ended in doing a MemTest86 try only after various other different tests, as I was not really thinking that it could throw me back any errors, as my ram is quite new and never overclocked over its specifications.....but I was contradicted!!

First of all my system:

OS Windows 10 Pro
Motherboard ASUS ROG STRIX Z370-G GAMING (WI-FI AC)
BIOS Version 1601
CPU Intel Core i5-8600K @ 3.60GHz
RAM 16GB (2x8GB) DDR4 XMP PC4-28800 G.Skill / Trident Z F4-3600C15-16GTZ 15-15-15-35 / 3602 MHz / 1.350V
Graphic Card Gigabyte GeForce GTX 1070 8GB Mini ITX OC 8G

State of overclock: CPU no overclock / RAM set at XMP parapeters / GPU no overclock

Now the story. It is a couple of weeks that the pc has some strange behaviors, like exiting from pc games, or bsods like these:

On Sun 10-Feb-19 9:18:46 PM your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\021019-8171-01.dmp
This was probably caused by the following module: ntoskrnl.exe (nt+0x1AA0A0)
Bugcheck code: 0x1E (0xFFFFFFFFC0000005, 0xFFFFF80159525388, 0x0, 0xFFFFFFFFFFFFFFFF)
Error: KMODE_EXCEPTION_NOT_HANDLED
file path: C:\Windows\system32\ntoskrnl.exe
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: NT Kernel & System
Bug check description: This indicates that a kernel-mode program generated an exception which the error handler did not catch.
This might be a case of memory corruption. This may be because of a hardware issue such as faulty RAM, overheating (thermal issue) or because of a buggy driver.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.


On Sun 10-Feb-19 9:18:46 PM your computer crashed or a problem was reported
crash dump file: C:\Windows\MEMORY.DMP
This was probably caused by the following module: ntkrnlmp.exe (nt!RtlCaptureStackBackTrace+0x62D)
Bugcheck code: 0x1E (0xFFFFFFFFC0000005, 0xFFFFF80159525388, 0x0, 0xFFFFFFFFFFFFFFFF)
Error: KMODE_EXCEPTION_NOT_HANDLED
Bug check description: This indicates that a kernel-mode program generated an exception which the error handler did not catch.
This might be a case of memory corruption. This may be because of a hardware issue such as faulty RAM, overheating (thermal issue) or because of a buggy driver.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.


On Sun 10-Feb-19 9:15:36 PM your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\021019-7875-01.dmp
This was probably caused by the following module: ntoskrnl.exe (nt+0x1AA0A0)
Bugcheck code: 0x50 (0xFFFFB80400000001, 0x10, 0xFFFFB80400000001, 0xC)
Error: PAGE_FAULT_IN_NONPAGED_AREA
file path: C:\Windows\system32\ntoskrnl.exe
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: NT Kernel & System
Bug check description: This indicates that invalid system memory has been referenced.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.


On Sun 10-Feb-19 8:59:33 PM your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\021019-8000-02.dmp
This was probably caused by the following module: ntoskrnl.exe (nt+0x14FB3E)
Bugcheck code: 0x1000007E (0xFFFFFFFFC0000005, 0xFFFFF8017C16BB3E, 0xFFFFC404E1BD26C8, 0xFFFFC404E1BD1F10)
Error: SYSTEM_THREAD_EXCEPTION_NOT_HANDLED_M
file path: C:\Windows\system32\ntoskrnl.exe
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: NT Kernel & System
Bug check description: This indicates that a system thread generated an exception which the error handler did not catch.
This appears to be a typical software driver bug and is not likely to be caused by a hardware problem.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.


On Sun 10-Feb-19 7:15:26 PM your computer crashed or a problem was reported
crash dump file: C:\Windows\Minidump\021019-7937-01.dmp
This was probably caused by the following module: ntoskrnl.exe (nt+0x1AA0A0)
Bugcheck code: 0xEF (0xFFFFC8852BFFB580, 0x0, 0x0, 0x0)
Error: CRITICAL_PROCESS_DIED
file path: C:\Windows\system32\ntoskrnl.exe
product: Microsoft® Windows® Operating System
company: Microsoft Corporation
description: NT Kernel & System
Bug check description: This indicates that a critical system process died.
There is a possibility this problem was caused by a virus or other malware.
The crash took place in the Windows kernel. Possibly this problem is caused by another driver that cannot be identified at this time.

Then today I started a game (FPS game) and after playing a while the game freezes up...then it takes a while for me to reach the desktop with several Alt+Tab and Ctrl+Alt+Del stuff...and the a nice BSOD! (one of those expressed up here)

After reboot the pc was ok, but I decided to update the motherboard drivers and any peripheral drivers that were not up to date, as reading at the reasons of that bsod on the net, it could have been something drivers related. So I updated my chipset drivers, my sound card driver and my sata drivers. All the other were ok

After all this I restarted the game again, and after a while it exited by it self, with its "report error" application asking me to send the developers the error log.

At that particular moment I remembered that this kind of error with the "report error" nag screen coming out was happening quite often in those last times I played at this particular game, such that I remember that I ended up in leaving it a side and dedicate my self in playing at another game as I had to finish it (older game less heavy in hardware min specs required). I though that the problem was that the game was faulty and soon an update would have come.

But just after remembering so, I remembered another thing!! That even at this older game I was having problems, like exiting once in a while like every 5 or 10 mins, depending on the load of the game....but I thought that was its last update that I installed about 10 days ago (since then I started in seeing strange exiting from the game) and never-minded as I was so close to the end that I decided to end it (yesterday) like that.

After all these reminding, as though that it could be the graphic card, I launched a GPU benchmark (using MSI Afterburner and its benchmark utility) so to stress the GPU a bit....and after a while I got another of the BSOD expressed up here.

At boot the pc was very slow in opening folders and moving files, that I though that all these bosds broke my Windows installation, but after a while it started to run as usual.

So I thought it could be the graphic card. But I needed to go deeper!! I ended up my browser and there was a YouTube page previously opened that started to play a video....and after a few secs another BSOD! (still one of those expressed up here)

At restart, I re-opened the browser and after a few secs that the video started to play...BSOD!! (again up here)

So I tried not to load nothing and to launch AIDA64 and test the CPU. So I choose to test CPU, Cache and RAM all together...and after a while AIDA stopped the test throwing me a Warning: Hardware failure detected!! ...never saw nothing like that!!! Tried only selecting the CPU...same thing! Selecting only the RAM.....same thing!!!!!

So I tried Prime95 with the TortureTest on mostly the CPU and low RAM...and the test ran for some time with no errors....so I tried the TortureTest that was using lots of RAM, and it stopped for ERROR FAILURE telling me that some random numbers were supposed to be some value, but eventually were others.

After all this I tried MemTest86, and in attach here are the results of pass1 and half of pass2, as I exited it as I was already at more that 50 ERRORS!!!!

Sooo.....what could it be?? Faulty RAM or CPU or what?? I would say that it is not the GPU.
About the GPU benchmark, I really don't understand how could it result in a bsod as the CPU was at barely 60% of load. Can't say nothing regarding the RAM utilization as I forgot to check with the Task Manager.

Please please please give me your advice. For the moment I have disabled the XMP for the RAM, so now they are running at their GEDEC default that is 2133MHz. Will try MemTest86 and see what it says!

In the mean time, please give me your thoughts. Thank you all!

[PIRATA!]

Hello again! Did the test at GEDEC ram defaults of 2133MHz, and so far no results after 2 and a half passes (then I ended it).

Any thoughts??

Please let me know.
Thank you.

[David (PassMark)]

MSI Afterburner had bugs in it that crash other software. See,
https://www.passmark.com/support/per...r_messages.htm
So far from being a diagnostics tool, it can actually cause problems.

This is the 2nd time today I have seen a MemTest86 log with errors from a combination of ASUS motherboard and GSkill RAM. In both cases there were errors in Test 6 and test 8. Both cases has a small number of bits in error over a wide memory address range. Maybe this is a co-incidence, or maybe not.

In that case turning off XMP fixed the problem as well. Which is a sign the timings or voltages in BIOS aren't correct. e.g. a firmware bug in ASUS BIOS, or bad SPD settings in the GSkill RAM sticks.

Your case
Intel Core i5-8600K
ASUS ROG STRIX Z370-G GAMING (WI-FI AC)
G Skill Intl / F4-3600C15-8GTZ 8GB DDR4 XMP PC4-28800

Their case
Intel Core i9-9900K
ASUS ROG MAXIMUS XI HERO (WI-FI)
G Skill Intl / F4-3600C17-8GTZR 8GB DDR4 XMP PC4-28800

If you have spare RAM is it worth swapping it out to see if it fixes the problem. You might also be able to manually get it working with very slight changes in voltage levels, timing changes & refresh rates. But this is something only experienced people generally play around with (and it should have been done by ASUS already when they setup the BIOS)

[PIRATA!]

Hello there! As usual, your help is precious.

Woww!! is something like a relief, what you just wrote. I'm sorry for the other guy, but at least I'm not alone.

The sad thing is that happened on a similar ram than mine, as the CL17 is quite the same, but with worst chips (if we can call them "worst"), so this is the confirmation that 1) these rams are quite tight, and 2) ASUS is a good brand but for geeks.

I've built this pc after been 10 years without a desktop pc, and as an ex geek my self (don't have the age for dedicating time in it any more) I went for what I remembered from the past...and maybe I should have gone stock, who knows

Yesterday I was also thinking in calling Amazon, as I bought the kit there "sold & shipped" by them, so to ask for refund or change with a new equal...but I read here that you say to try first to swap the banks or play with timings and voltages.

I was even hoping that the rams were broken, so to eliminate the problem...but from your words I read that these things could be possible with kits and motherboards like these.

Now that I recall, I might have to say that had no problems of these kind with the previous motherboard bios, as since when I updated it in November..well I recall these problems! At beginning were just a few bene exits, then the first bsods..then you know already from my previous post.

Should I roll back to the previous buib I was using before (if it's possible to do so) and see if it fixes the problem, of should I first swap the banks, or play with the settings??

Ps: I'm not that good any more with all these super complicated bios with all these settings. Do you know which are the settings to look after in the bios, a part from ram voltage?? (that by the way my motherboard shows to give even slightly higher V. than those requested and set in the SPD by G.Skill -- I have a little box showing the SPD Volts by G.Skill, and next to it there is a box indicating what is the motherboards giving to the Rams...and this is slightly higher)

....or would you go for calling Amazon directly?

Pps: thanks also for the hint on MSI Afterburner benchmark...but I have to say that in the past I tried it and don't recall in having no issues (used same benchmark version then and now)....and I was with old bios!

Thank you for everything..as usual

[David (PassMark)]

I was even hoping that the rams were broken

Well something is broken. You should just plug in the sticks and they work.
But to really work out scientifically what the problem is you need some high end oscilloscopes and bus analysers. ASUS and GSkill should be doing this work. Otherwise it is guessing, trail and error.

There is a lot of marketing pressure on the vendors to offer more performance at lower prices. Which means tighter timings, less compatibility testing and less tolerances for slightly marginal parts.

XMP doesn't in fact offer a great deal of extra performance for the system as a whole. So you can likely turn it off and not notice. But in that case you've basically bought high end parts and can never use their full performance.

If it was me I would try out the suspect RAM in a few different motherboards and try some different RAM in the suspect motherboard. But you need to have a lot of high end spare parts sitting around to do this.

[PIRATA!]

Heh...it's the same old story: hardware makers that makes all the super tests cross-platforms for compatibility proof?? I lost hope for something done right!

I don't know if the ram producers tests really each motherboard they write in their compatibility lists (and mine is precisely specified in my ram QVL), and if the do so, they do it on the same motherboards used all over for each new ram of that window of time of the existence of those motherboards, as they have to specify them on different QVLs of different kind of rams. So I imagine that those reference motherboards are pretty well used, and that could easy give back different results from some other brand new motherboard of the same model. And we are considering here that each same motherboard should perform the same, that it's something only theoretically, as electronic components (specially on these well extreme motherboards) can perform and behave differently. Some times a production branch could perform differently from one to another... Can't imagine a test motherboard heavily used compared to a brand new motherboard assembled at months of distance in production dates...!!!

Speaking of my case, I have swapped the ram sticks, and this time no errors at test6 but only on test8 (16 errors so far)...but when I went for existing and save the report...the pc froze, as well as MemTest86, leaving everything still on monitor...no auto-reset or blank screen or distorted image: just all frozen with the Esc menu popped up...so that I had to hit the resre button!!

Going back to take a look at the voltages, I realized that even if the Auto voltage was set at SPD, the one set by the motherboard was V1.3530, and that the voltage monitor from the motherboard's bios was saying V1.344.
So I thought that either the motherboard was not able to follow the correct SPD setting, or that the monitor was not reading the correct voltage...or both!

For making a test, I have set the voltage at V1.3540, that the bios rounded to V1.3592 (can not set something in between...), and in the monitor is shown as V1.360.

So I'm now running MemTest86 with these voltages as I write, and I have done so far only one pass and a half, and had only one error in test8.

Soo...a voltage error??

As I'd like to learn more about which should be the correct voltages of all each components or circuits of such complicated computer (as I can change any voltage here, even circuits voltages), do you know where I could find some useful informainformation?? I've made some researches but haven't found nothing good (I'm speaking about something like: "this should work at this V but can tolerate these +/- V...and this can work freely in this range of V...")

I say this because before this last bios update (which I still think it could be one of the culprit) I had to play around the cpu voltages and related circuit's voltages for setting what the correct voltage should have been, as the "Auto" was not serving goods Volts. Now with the new bios update, my motherboard is giving all the "correct" voltages when setting everything to "Auto"...but I'm really suspecting that the voltages I see given by the motherboards are not what they are supposed to be!!!

For all just sayd, I am suspecting that maybe Asus have worked on the voltages in the last bios update...and this have messed up the ram stuff, so to compromise the already critic situation of a pumped motherboards asked to work flawlessly with pumped ram.

Ps: I don't have another motherboard to test my ram on...but I can buy another ram kit just same as mine, and check if they behave the same, and maybe mix them so to find a working solution. Would you try that..or what??

Please let me know, within your thoughts, that are always widely appreciated.

Thank you.

EDIT: here in attach is the result with only 1 error.

[PIRATA!]

Made another test: setting Voltage back to the XMP defaults, so V1.350 (that in the motherboard monitor is read like V1.344) and set timings to 15-15-36 (instead that 15-15-35):

- errors on test5 and test7 !!!!

[PIRATA!]

Made also another test: same XMP stock Voltage of V1.350 (V1.344 in motherboard monitor) and set timings to 16-16-36 (instead that 15-15-35): No Errors at first pass!!!!

[David (PassMark)]

So maybe your expensive C15 RAM should have been better tested by vendor and sold as C17 RAM. Or maybe ASUS motherboards aren't ramping the timing signals fast enough, which is why more voltage helps a bit. Maybe a combination of both.

[PIRATA!]

So maybe your expensive C15 RAM should have been better tested by vendor and sold as C17 RAM. Or maybe ASUS motherboards aren't ramping the timing signals fast enough, which is why more voltage helps a bit. Maybe a combination of both.

Yeah...

Could it be that these rams have ended their "working" period and now they work worst than before? Meaning that my motherboard is not the problem? I mean...what is it more possible here: the ram are in fault, or the motherboard is??

Second thing...rolling back to previous bios would it be something you would try??

Thanks!!

[PIRATA!]

Ok, just contacted Amazon and they are sending me a new kit as a replace. I'll have them both for some time, as I have 1 month for sending back the faulty one, so I'll make some tests.

In the mean time a new bios update for the motherboard was been released, so I'll test that as well.

I'll update you guys!!

[PIRATA!]

Updated bios to last version. Set back everything to XMP active like before.
Now I don't have any errors on test6, but in test7, test8 (most of them here), and test9. Then quit test.
In attach result.
Any suggestions by reading the address range??

Will wait the new kit for further tests.
Thank you.

[PIRATA!]

So maybe your expensive C15 RAM should have been better tested by vendor and sold as C17 RAM. Or maybe ASUS motherboards aren't ramping the timing signals fast enough, which is why more voltage helps a bit. Maybe a combination of both.

Hi again!
Ok..Amazon sent me a replacing kit of exact ram...and they are faulty as my original one!! Here in attach the two tests that I've run on the new kit (the second test is made with the sticks switched).
Both tests have failed quite fast mostly on test8....does this give you any clue??
Could it be a motherboard settings problem? Like some CPU variables or voltage variables??

Tomorrow I'll try to mix up the sticks between the two kits, but I will not be able to keep any of those combination because I can't mix up the kits and will have to return the faulty kits as they were shipped....but will do the tests for seeing if that could be a solution.

QUESTION: if I'll decide to go for returning everything and go for another kit...as you know, what are the most stable type of rams extra Jedec's limit?
I mean..Jedec says that the limit is 2666 and that higher are OC...but which one of the OC's are those with less reported issues as you know? 2800? 3000? 3200?
And Speaking of CL? If I would like to focus on the less probable to be faulty, is it better to go for a stated 3600 CL17 or a 3200 CL15??

Thank you.

[David (PassMark)]

For the minimum risk of problems consult the qualified vendors list (QV list) for the motherboard.

List is here for your motherboard,
https://dlcdnets.asus.com/pub/ASUS/m...ING_190218.pdf

They have the F4-3600C15D-16GTZ on the list, but not F4-3600C15-16GTZ. Not sure if these are the same model RAM or not?

Otherwise I don't think we have anything new to add, I'd just be repeating what I already said in previous posts.

[PIRATA!]

For the minimum risk of problems consult the qualified vendors list (QV list) for the motherboard.

List is here for your motherboard,
https://dlcdnets.asus.com/pub/ASUS/m...ING_190218.pdf

They have the F4-3600C15D-16GTZ on the list, but not F4-3600C15-16GTZ. Not sure if these are the same model RAM or not?

Otherwise I don't think we have anything new to add, I'd just be repeating what I already said in previous posts.

Heh..I have the F4-3600C15D-16GTZ. Yes that are mine. I forgot to put the letter "D" in their name in the first post. Before buying the ram kit I checked in the QV list of my motherboard, and as you can see that does not mean that it works.

As I don't have another motherboard to test them, I have to make my tests with what I have, so I have to go for some more generous kits.

Was asking if you know in your experience if there are some more centered Frequencies/CLs that are more reliable that others...so to check in my QVL if I can find something....so if maybe a PC3000 is historically more reliable than 3600...or if it is more a matter of CL where higher CLs are more reliable that lower CLs, regardless the Frequencies...

And regarding the kit composition, is it true that for example on a 32Gb kit, a 16x2 is better that 4x8?? Or is it the opposite...or a myth??