Game frozen and them crashed to windows

[Kunish]

8 minutes ago, Beelzybub said:

@Kunish

You seem to have gone through all the software related possibilities, and I'm convinced at this point that you

have a hardware/driver problem. I would root through the Windows Event logs to see if there is a new clue there.

Maybe the Windows 10 Reliability Monitor will have a clue in it.

 

Some of your errors suggest a disk problem. I know it passed the tests though so it's hard to say what's up.

It seems to be the only clue we have to go on at this point.

Maybe have a look at the bios and see if there are any options for the nvme drive.

Is it running in pcie v4 mode? Can you switch it to pcie v3 mode? I dunno really. Are there any network drives, or

iSCSI drives?

 

Sometimes a system builder will accidentally leave the plastic on the thermal pad of the nvme heat sink, which

could cause errors.

 

An extreme test would be cloning the nvme to another drive, (an ssd) and removing it.

I installed the nvme, no plastic , but I was thinking in installing windows in the other ssd and try again, the new components of my computer are motherboard, processor, memories and the nvme, I hope is not a hardware problem, but I think it have to be a windows problem or maybe some incompatibilty between the nvme, windows and the game. This week-end I´ll format the nvme and do a new install of windows, if that not work I´ll remove the nvme and install windows in the ssd, wish me luck.

[Jugginator]

Diskspd.exe -b8K -d60 -h -L -o2 -t4 -r -w30 -c50M c:\io.dat

 

You can try that out, replace c:\ with the drive letter of yours.

 

There is an open-source disk stress-tester called crystaldisk, it's based off of Diskspd (that's a built-in Windows function).

 

But... yeah I mean, I really second Beelzy in this may be a drive failure somewhere, or something of that nature. Conjecture, but SSD's aren't as long-lasting as HDDS either (if you have an HDD that has great mechanical parts, that is).

[Beelzybub]

8 hours ago, Kunish said:

the new components of my computer are motherboard, processor, memories and the nvme

When you installed these new parts, did you keep the existing Win10 install, or start fresh?

 

7 hours ago, Jugginator said:

(if you have an HDD that has great mechanical parts

Off topic but, one of my crap test machines has an 80 Gb HDD from... 2002? I think?. Runs fine, reports no smart errors or block errors. One of my bug reports acually was recorded on that drive. It won't die, and as long as it's working, I can't bring myself to throw it out.

[Jugginator]

57 minutes ago, Beelzybub said:

When you installed these new parts, did you keep the existing Win10 install, or start fresh?

 

Off topic but, one of my crap test machines has an 80 Gb HDD from... 2002? I think?. Runs fine, reports no smart errors or block errors. One of my bug reports acually was recorded on that drive. It won't die, and as long as it's working, I can't bring myself to throw it out.

Yeah right, if you get a good HDD it will last you forever if you don't treat it rough.

[Kunish]

20 hours ago, Jugginator said:

Diskspd.exe -b8K -d60 -h -L -o2 -t4 -r -w30 -c50M c:\io.dat

 

You can try that out, replace c:\ with the drive letter of yours.

 

There is an open-source disk stress-tester called crystaldisk, it's based off of Diskspd (that's a built-in Windows function).

 

But... yeah I mean, I really second Beelzy in this may be a drive failure somewhere, or something of that nature. Conjecture, but SSD's aren't as long-lasting as HDDS either (if you have an HDD that has great mechanical parts, that is).

These are the results.

Diskspd.txt

20 hours ago, Beelzybub said:

When you installed these new parts, did you keep the existing Win10 install, or start fresh?

 

Off topic but, one of my crap test machines has an 80 Gb HDD from... 2002? I think?. Runs fine, reports no smart errors or block errors. One of my bug reports acually was recorded on that drive. It won't die, and as long as it's working, I can't bring myself to throw it out.

A fresh start in the nvme.

[Jugginator]

Your nvme is fast.

 

Really the only two concerning things are your cpu and kernal worked rather hard for the test and:

 

5-nines |     11.950 |      1.232 |      7.559
6-nines |    464.266 |      4.358 |    461.062

 

That's a huge jump in time. Normally it's a little more progressive. If a clean install doesn't work... I am kinda sorta leaning to it (drive issue) but not too sure yet.

[Beelzybub]

35 minutes ago, Jugginator said:

That's a huge jump in time.

Maybe it went from cached data to un-chached data at that point?

[Jugginator]

9 minutes ago, Beelzybub said:

Maybe it went from cached data to un-chached data at that point?

Yeah but still that's a large jump. I ran it on my m,2 and it jumped about 80-90, not over 400

[Kunish]

Ok, I did a clean install and the game crashed again, I played for about 40 minutes in navezgane, this time I thinked that the problem was resolved, no configuration changes in the game, I only install the amd driver, steam and 7 days to die, the crash to windows desktop was faster than ever, even in the outlog the crash is not registered, I attach the output log and the event log. My only solution now is install another clean installation of windows in another disk and if this fail I´m out of options.

output_log__2020-09-17__22-24-06.txt event log.txt

[Beelzybub]

2 hours ago, Jugginator said:

Yeah but still that's a large jump. I ran it on my m,2 and it jumped about 80-90, not over 400

That could be from a MLC vs. TLC vs. QLC thing. A QLC drive would have a huge jump.

[Kunish]

Update: I have managed that the game not crashed, I disabled the xmp profile in the bios, and after that no problem with the game, them I did several test for the RAM because I supposed that I had a problem with it, but every test was OK, I even test the RAM with memtest, 6.5 hours, and zero problems, I don´t know the cause of the crashes, the profile that I had was one that left my RAM at 3000 and CL15, I have to try with the other profile, 3200 and CL16 and see if this profile also crashes.

[Jugginator]

Sweet!!! I'm glad to hear you've progressed, I was wondering how you were going with it. Disabling the xmp profile leading to better results almost makes me feel there are minor faults on your motherboard that aren't glaringly obvious but cause instability with heavy read/write load under the xmp profile... yeah?

 

I know when messing with my timings on my ram (Ryzen cpu) and also increasing my OC very slightly more (separate times) I would be fine in stress test, fine in 7 Days and crash 3/4th of the way through in Cinebench (or, fine in stress test, fine in Cinebench, and crash in 7 Days after 20 mins). So, I leave my settings on my best stable setting now. But XMP works fine.

 

Maybe your motherboard is overheating?

[SylenThunder]

1 hour ago, Jugginator said:

Sweet!!! I'm glad to hear you've progressed, I was wondering how you were going with it. Disabling the xmp profile leading to better results almost makes me feel there are minor faults on your motherboard that aren't glaringly obvious but cause instability with heavy read/write load under the xmp profile... yeah?

 

I know when messing with my timings on my ram (Ryzen cpu) and also increasing my OC very slightly more (separate times) I would be fine in stress test, fine in 7 Days and crash 3/4th of the way through in Cinebench (or, fine in stress test, fine in Cinebench, and crash in 7 Days after 20 mins). So, I leave my settings on my best stable setting now. But XMP works fine.

 

Maybe your motherboard is overheating?

Yeah, I kind of agree with this. Typically the fix in regards to memory timings is to set the XMP profile, because it runs underclocked/unstable without that set. 

[Kunish]

19 hours ago, Jugginator said:

Sweet!!! I'm glad to hear you've progressed, I was wondering how you were going with it. Disabling the xmp profile leading to better results almost makes me feel there are minor faults on your motherboard that aren't glaringly obvious but cause instability with heavy read/write load under the xmp profile... yeah?

 

I know when messing with my timings on my ram (Ryzen cpu) and also increasing my OC very slightly more (separate times) I would be fine in stress test, fine in 7 Days and crash 3/4th of the way through in Cinebench (or, fine in stress test, fine in Cinebench, and crash in 7 Days after 20 mins). So, I leave my settings on my best stable setting now. But XMP works fine.

 

Maybe your motherboard is overheating?

Thanks,

I didn´t see anything strange in the maximun temperatures when the crashes happened, my motherboard is an Aorus x570 ultra that have a generous heatsink https://www.gigabyte.com/us/Motherboard/X570-AORUS-ULTRA-rev-11-12#kf , I have also a ryzen (3700x) and the crashed happened in default and also crashed with an underclock at 3900mhz and 1.175 volt, I have to try the xmp at 3200 and see what happen, in any case without the xmp I´m only lose about 5fps and at least I can play

[Kunish]

Ok, huge surprise, the game appears to be stable with the other xmp profile, 3200 cl16, the profile that gave me all the problems left the RAM at 3000 cl15, It seems that put the RAM at cl15 cause some kind of problem with this game, at least with my configuration, I hope that this help another one that have a similar problem, thanks to all, if I have some future problem relate to this I´ll update again.

 

Btw, my RAM is this: Kingston HyperX Fury Black 16GB DDR4 3200Mhz PC-25600 CL16

[meganoth]

1 hour ago, Kunish said:

Thanks,

I didn´t see anything strange in the maximun temperatures when the crashes happened, my motherboard is an Aorus x570 ultra that have a generous heatsink https://www.gigabyte.com/us/Motherboard/X570-AORUS-ULTRA-rev-11-12#kf , I have also a ryzen (3700x) and the crashed happened in default and also crashed with an underclock at 3900mhz and 1.175 volt, I have to try the xmp at 3200 and see what happen, in any case without the xmp I´m only lose about 5fps and at least I can play

XMP profiles only have a small impact on heat. What they change is the timing of how the RAM is accessed.

 

When an XMP profile gets one of its timing intervalls too close to a limit, it can result in some reads or writes to or from RAM not succeeding and simply producing garbage.

It would mean the machine is still operational, but either because of random fluctuations of timings or because in a specific situations one of the signals arrives minimally later the machine reads or writes garbage and the machine crashes.

 

This also means it could be the slowest profile that has a problem and the fastest being the best (although in general the fastest profile should have the smallest safety margins)

 

 

[Jugginator]

6 hours ago, Kunish said:

Ok, huge surprise, the game appears to be stable with the other xmp profile, 3200 cl16, the profile that gave me all the problems left the RAM at 3000 cl15, It seems that put the RAM at cl15 cause some kind of problem with this game, at least with my configuration, I hope that this help another one that have a similar problem, thanks to all, if I have some future problem relate to this I´ll update again.

 

Btw, my RAM is this: Kingston HyperX Fury Black 16GB DDR4 3200Mhz PC-25600 CL16

Super happy to hear you can finally play the game again, good work sticking with it!! Yeah, bad timings/clock on ram can cause a lot of instability. 7 Days definitely will show you those issues LOL.

On 9/11/2020 at 4:15 PM, Kunish said:

Two days ago I did several stress test(occt)and everything was ok.

 

In my recent try, with the new drivers, I played from about an hour and the game didn´t crash, but I had to left the game from a very strange bug, I couln´t move the player with the mouse, I imagine that this bug is not related with the crashes, I try load the game several times but the bug reproduce in a few seconds. Tomorrow I´l try something, thanks again. I left a video of the bug: https://streamable.com/ti18pt

Now that bug if you see it again let us know too. Very weird one.