Gigabyte GTX 1060 3GB. Games often crash. What troubleshooting can I do to help fix it?

[Danny75]

Annoyingly, I managed to play Ghost Recon for over 2 hours without issue.

Is that normal? Any signs that the CPU voltage increase has diminished the rate of crashes? Or can't tell one way or another?

However, I will still change those settings and see how it goes.

Just to add, I tried the graph but when I dragged the point down and hit apply, it sprang back up to its default.

I think I figured it out, I had to ctrl drag to move them all, otherwise it would have a single drop and I don't think it liked it.

Nice. So when you do that, it'll be locked to that frequency and voltage all the time. It's fine for a while but it's just to test for stability. At some point just unlock it by selecting the same square in the curve and press L again. Or just hit the Reset button in Afterburner and apply.

 

[X82]

Here is a link to the new hardware monitoring file. This was done after the voltage increase but before I locked the voltage etc.

On that note, being able to play for long periods is sometimes normal, which as I said, frustrates. Before I made any changes last night, I managed to play Arma 3 for just over 3 hours without issue.
But I by no means take it as fixed. In my book, if I can go through a week with no crashes, its fixed.

About the locking of voltages, I noticed when I hit apply after locking it, the numbers didn't change. Do they only when I play a game?

 

[Danny75]

Had me stumped why the voltage and frequency didn't change for you, then I remembered Voltage Control and Voltage Monitoring need to be unlocked from Settings > General tab in Afterburner. Sorry.

Had a look at the new log you uploaded as well. Can't see anything odd there either.

 

[X82]

I was getting worried as that was already checked. But a restart of the application fixed it.
Running now at 1708mhz and 1000mv. Let us see how this performs over the next few days.
Again, thanks so much for putting the time in to give me some thorough answers.

I will report back, but I do imagine if it keeps crashing, my options are severely limited.

 

[Danny75]

I was getting worried as that was already checked. But a restart of the application fixed it.
Running now at 1708mhz and 1000mv. Let us see how this performs over the next few days.
Again, thanks so much for putting the time in to give me some thorough answers.

I will report back, but I do imagine if it keeps crashing, my options are severely limited.

Post back whenever you want and don't worry about the time, it's like a hobby.

Did mention breadboarding it outside of the case. That's really the first thing a computer repair person should do, if being thorough, to eliminate any shorting, unless they suspect a particular component and can swap it out for another.

What about GPU sag, is there any? Even if there isn't any major sag, might be worth a try to lift up/straighten the card slightly more just to see. Unscrew it from the PCIe brackets and try to screw back in with the card a millimetre or two higher. Blow any dust out from the slot while you're at it.

Just some ideas ^

 

[orbitalwalsh]

blck... able to drop it down to 100? might be a little to much for pcie slot - then again been a while overclocking with older intels and blck clocking

 

[X82]

Well, since I last posted life got in the way so only just got around to testing Ghost Recon.
Played for about 30/45 minutes before crash again.
New monitoring file here.
This is after locking it to those new values.

So yeah, square one?

 

[Danny75]

Well, done what you can unless any other ideas are suggested. You could RMA the card and tell them everything you've tried. If they find no fault, or the replacement does the same at least you'll be more confident that the issue lies elsewhere. As you said could be dodgy motherboard, especially if it happened to other GPUs. Or PSU flaking out.

Bit of cheap maintenance would be to replace the CMOS battery (should be a CR2032 lithium coin but check first). Unlikely to be the issue here but should be done anyway given the age of the board.

Remember to reset Afterburner so it unlocks frequency/voltage.

 

[Unseul]

Take the side off the case, run fans at 100%?

Very different cards, but I experienced crashes with my rx580, and it seems likely it was due to spikes in temp, since when I'm running everything full whack, no more crashes. Normally the card temps would be 60s-70s, but i guess if it spiked it went to high. Now it doesn't break 70 and I've not experienced any more crashes.

I know you're keeping logs of it, but it's worth a shot. Completely uneducated suggestion btw, just what helped me.

 

[LuckyBenski]

I really think this is motherboard or power supply related, as it's always done it.

You don't need a Z68 board because you were never overclocking and you haven't had working USB 3. I'd find an H61, H67 or P67 board. Test that and a new PSU - only needs to be 400W maximum for your parts.

 

[X82]

Firstly, thanks to all who replied, some great ideas and gave me places to look.
I need to be logical here, somewhat at least. My games crash, but it isn't a constant thing. It's annoying but I've lived with it for years now.

I really think this is motherboard or power supply related, as it's always done it.

You don't need a Z68 board because you were never overclocking and you haven't had working USB 3. I'd find an H61, H67 or P67 board. Test that and a new PSU - only needs to be 400W maximum for your parts.

After all these years I'm still being schooled. I wrongly assumed it was the more expensive boards I needed, I couldn't afford the £100+ for an old board which MIGHT fix my issue. So thanks for that information.
Since the boards I've looked at in the right price have only 2 memory slots, I will drop 8GB ram if I go down that road. Getting a 4 slot board is an option, but those are made by Foxconn and I have no idea if they are a good brand.

Moreover, if I buy a new motherboard and PSU and it does fix this issue, I wouldn't have gained too much. Does that make sense?
People buy components for their machines for a boost. I will still get the same **** poor frames in some games.

I suppose I have to weight it up:

Functioning USB ports
No more crashes (possibly).

Over the random occasional crash and one functioning USB port.

Most importantly, there is no guarantee that replacing those parts would change anything. I could play a game for hours and no issue and then a crash.
Did it crash because sometimes games just crash?
Or was is another issue. Memory? CPU?

I need a thorough upgrade and I am hoping if my finances play nice, I have a plan hopefully at the end of the year to buy an entirely new machine and I will no doubt be calling on you guys for pointers and assistance during that period.

So again, thanks for the input. I think since last night I managed to play Arma 3 with no issues, I will just put up with this annoyance until I can afford to go big or go home.

 

[X82]

Just thought I would post a small update.

I originally said above I can put up with it. But it is starting to get to me now.
Thanks to the Steam summer sale, I bought a couple of games. Games I thought were not too taxing but so far they have both crashed with the same freezing and then hello desktop.

The system does bluescreen when just sitting idle too with each reason for it being different the I have no idea what I am looking at since googling the bluescreen codes brings up a wealth of possible answers.

I do wonder if it is the memory, but I can't afford a new pair to even test it.

This is what my bluescreens look like currently:



I do think there is little much I can do. Old board and old CPU.

 

[Danny75]

You said you'd tried Memtest and it passed. Have you tried Memtest HCI as well? https://hcidesign.com/memtest/

 

[X82]

I haven't no. I will run it over night and see if it finds anything.

 

[decto]

Bit of a long shot, but there was a bug in the M4 SSD's that caused a crash every hour or so once the drive had 5000ish hours of use.
I'd assume you'd have patched this years ago, it needed a firmware update.

As you seem to have covered all the other bases, and you mention it happened with other GPU's may be worth checking you have the latest firmware version.
https://www.crucial.com/usa/en/support-ssd-firmware

 

[X82]

Thanks for that.
I did have issues with the SSD fairly early on and I updated the firmware. That was in 2013 and there hasn't been an update since it seems.

 

[X82]

Just thought I would post an update which may point a finger to the problem.
Firstly, I've yet to run a memtest. In truth, I keep forgetting each night.

I've noticed however that for a lot of my games when they crash and when they do display an error message, it is always some Access Violation. Arma 3, Satisfactory, Ghost Recon and now FAR - Lone Sails.

Usually after this (although not always) I get a blue screen.

Here is a collection of all the minidumps in case it is useful.

 

[Quartz]

Every single game I play will crash after a duration.

What does the Event Viewer show? The very first thing you should do is check the Event Viewer. Look for entries with red icons.

 

[X82]

Well, the frustration is growing.

I've ran memtest, all clear
Nothing in event viewer of note.
All HDs scanned and checks passed.
Increased CPU Vcore
Even going so far as to disable Data Execution Prevention (DEP). Games still crashed

The only pattern I have is every single game that crashes, reports in some form or another (in its own log files) that it crashed due to EXCEPTION_ACCESS_VIOLATION.
I've Googled for hours to solve this and tried everything I can.

As I've mentioned previously, I think it will be a case that the only fix is to bin this POS a buy a new PC. But since I can't do that right now, I was literally praying for a solution.

 

[Danny75]

I've ran memtest, all clear

Memtest HCI as well? The thing about memory tests is that they are great when they tell you there are errors because job done, culprit identified. But they aren't always capable of finding them, and this results in failures during other operations. Not all that common but it's happened to me once. Only replacing the RAM did the trick. So unfortunately you cannot exclude that unless you test different memory (or find another culprit first).