[SOLVED] Is this used GPU a dud?

[Robbom]

Got a used msi gtx 1060 6gb ocv1 used off of ebay a few days ago and had trouble setting it up at first, as it kept working for a short time, then suddenly stopped producing a signal and i was forced to power the pc down
Ended up purging my drivers with DDU and then installing the GPU once again and after installing the correct drivers I've been using it all day (around 8 or 9 hours) with no issues, having played a few games and heaven benchmarked it on multiple settings with good results.
Suddenly when under pretty much no significant load (i just had a few browser windows open) the same black screen issue happened again and i had to shut the pc down
After now rebooting the pc i opened up msi afterburner as quickly as possible to see that while temperatures were alright, the clock speed was fluctuating quite heavily between roughly 130, 900 and 1300 mhz
The GPU is now just sitting at a low clock speed and i haven't tried running anything significant yet as I'm not really sure what my next steps are and could do with some advice on whether i should get a return for the card

My specs are:
Intel - Core i5-3350P 3.1 GHz Quad-Core Processor

MSI - B75MA-E33 Micro ATX LGA1155 Motherboard

Corsair - XMS3 8 GB (2 x 4 GB) DDR3-1600 Memory

Corsair 500gb SSD

MSI GTX 1060 6gb OCV1 (used to be MSI - GeForce GTX 660 2 GB Video Card)

NZXT - Source 210 (White) ATX Mid Tower Case

Corsair 500W PSU

Any help would be much appreciated

EDIT i just tabbed into afterburner again to see the clockspeed at 1400+ with memory speed at around 4000 with just 2 browser windows and discord open, this lasted for about a minute before returning to a lower value again

 

[Dark Lord of Tech]

I would return the card if the seller has that policy in place , not all do.

 

[Achaios]

GPU behavior is normal. The card however isn't normal, it's faulty. Return it if you can.

 

[Robbom]

Ok thanks, I have messaged the seller to get his take on the situation
Would/could this problem not occur due to PSU undervolting or motherboard bios incompatibility or other separate issues?
The driver I was running at the time of the crash was also around a year old and I've since upgraded to the most recent one could this have effected it?
Or is the card most likely the cause?

 

[MetallicMonk]

Ok thanks, I have messaged the seller to get his take on the situation
Would/could this problem not occur due to PSU undervolting or motherboard bios incompatibility or other separate issues?
The driver I was running at the time of the crash was also around a year old and I've since upgraded to the most recent one could this have effected it?
Or is the card most likely the cause?

I'm going through a similar problem currently. It could be a PSU issue but usually power related crashes can be reproduced ie: not random.
An old driver wouldn't cause damage to your card. Motherboard incompatibility I'm not sure about. The fact it works for a length of
time suggests that your motherboard is probably OK.

I would try uninstalling MSI afterburner and using the stock NVIDIA software to do any modifications to the clock/mem settings.
If you have only had two crashes in a few days I wouldn't be so hasty to return the card. GPU and memory clocks fluctuate depending
on load so that isn't a worry.

Also next time it crashes try ctrl+shift+windows key+B . This is the shortcut to reload the graphics driver. It can help avoid you rebooting.

It definitely warrants further testing.

Does the card have coil whine?
Try putting the power limit of your card to its max value. (not core/mem) It may be running out of power when pushed hard.
More likely would be a temperature issue. Use an overlay to monitor temps in game and during benchmarks and adjust your fan to a higher speed curve.

 

[MetallicMonk]

Ok thanks, I have messaged the seller to get his take on the situation
Would/could this problem not occur due to PSU undervolting or motherboard bios incompatibility or other separate issues?
The driver I was running at the time of the crash was also around a year old and I've since upgraded to the most recent one could this have effected it?
Or is the card most likely the cause?

Oh also install GPUz and make sure your BIOS is the same version as on the techpowerup website. The seller could have flashed a custom BIOS for mining etc

 

[Robbom]

Thanks for you replies
Checked the bios and it seems to be the correct bios for the specific card, also the seller says the card has been used on a family pc and he has never flashed any bios
I have increased the power limit to 116% on msi afterburner (couldn't find the nvidia software you mentioned in the brief time i had over lunch, sorry) and will try running some games and benchmarks tonight
Not really sure if i can even recreate the issue though to be honest as like i said the crash yesterday occurred under what should have been very low load and as far as i know the temperatures were fine (quickly rebooted and checked them and temperatures during gaming and benchmarks were monitored and were at max 50-55C)
Oh and in regards to coil whines, i have noticed there is a small noise for about half a second when my screen is turned on and off and when i alt + tab out of a GPU intensive programme or game in full-screen that sounds like a beep or whine, but i am not sure if this is what you're referring to
The seller has said he is happy for me to keep looking into this as long as i keep him posted as he hopes i can solve it, but also says that i can return the card at any time if i believe it to be faulty so at least theres that!
Anything else you can think of that i should look into? Is it possible that malware is responsible for this?

 

[MetallicMonk]

No problem,

Checking temps after rebooting won't tell you much, temps can decrease rapidly in a few seconds.
The coil whine you mentioned shouldn't be of concern at all, seems very slight and intermittent.

That being said, max temps of 50-55C during games are not high at all and shouldn't be the problem.

Also, since you have updated the driver the problem may have gone for good. edit; Forgot you had a crash after the new driver
Malware shouldn't cause problems like that, however there's always a chance.
Make sure you have windows power settings set to performance also.

Let us know the results of your benchmarking and games session.

I would run a stresstest as opposed to a benchmark. See if you can run OCCT or uniengine heaven for an hour or so.

Good luck

 

[Robbom]

Gaming session went well with no real issues, the one issue I did have was a weird thing when I went to exit the game where the GPU usage stuck at 1-5% and sort of froze the game up a few times. Couldn't really explain why or how this occurred to be honest, temps were fine no coil whine and it should have been under low load. It was weird that it happened at this specific point when I had been gaming under higher loads for much longer beforehand.
As for the drivers no you were right, the crash happened on the older drivers that were around a year old. I haven't had a crash since updating (note that the crash occurred after the pc was used for about ,5-6 hours whereas I gamed for probably 2:30 last night).
Another thing I found is that windows power settings had been set to balance before so they have been changed.
I'll run that test tomorrow, have another evening of gaming and see what happens. I think I'll have to make my decision whether to return or not then though as I don't want to mess this seller around too much. I'm leaning towards returning the card and forking out the bit of extra cash for a new one if I can't find a solid reasoning and solution to what has happened.
Thanks once again for your help

 

[MetallicMonk]

Gaming session went well with no real issues, the one issue I did have was a weird thing when I went to exit the game where the GPU usage stuck at 1-5% and sort of froze the game up a few times. Couldn't really explain why or how this occurred to be honest, temps were fine no coil whine and it should have been under low load. It was weird that it happened at this specific point when I had been gaming under higher loads for much longer beforehand.
As for the drivers no you were right, the crash happened on the older drivers that were around a year old. I haven't had a crash since updating (note that the crash occurred after the pc was used for about ,5-6 hours whereas I gamed for probably 2:30 last night).
Another thing I found is that windows power settings had been set to balance before so they have been changed.
I'll run that test tomorrow, have another evening of gaming and see what happens. I think I'll have to make my decision whether to return or not then though as I don't want to mess this seller around too much. I'm leaning towards returning the card and forking out the bit of extra cash for a new one if I can't find a solid reasoning and solution to what has happened.
Thanks once again for your help

Glad you haven't had issues. The issue exiting the game doesn't sound to me to be related to the card. More likely windows being temperamental. As soon as you exit a game the GPU usage will drop like that normally. You have a fairly old system like mine so shouldn't be concerned over small things like that.

All in all with what you have described I would say your card is 100% fine and I would keep it. But only if you are able to run uniengine heaven non-stop for an hour or so. That will be a good stress test for it.

You can always expect crashes from installing new hardware until drivers are sorted out and the fact you haven't had a crash since the new drivers makes me think you really don't have a problem.

Give the card a solid stress test on an infinite benchmark like heaven. If it passes, be happy and know your card is fine

 

[Robbom]

What settings should I run Heaven at? 1080p Extreme?

 

[MetallicMonk]

Yeh that sounds good. Whatever your monitor native resolution is.

 

[Robbom]

Well, i didn't even get to the test the card yet as it lost signal during three successive boots, firstly at the log in screen and the second two times before the motherboard info came up :/
Computer is completely unresponsive during this and trying to reboot drivers did nothing, however there is no coil whine noise and it seems the computer is acting as if it is still on hardware wise (all fans running power light on)
I also tried different HDMI cables during this and had no luck
At one point my monitor actually put itself into sleep mode which was weird, but i didn't think to try another monitor at that time and will try that if the problem happens again (imagine if it was just my monitor! haha)
I'm going to stress test now anyway and see what happens, mostly because i want to see if i can replicate the problem and try a few more things
Clearly though, the problem hasn't been solved just yet

---EDIT--- 45 minute Heaven benchmark at max, no problems

 

[MetallicMonk]

Oh that sounds annoying. Was this after updating any drivers?
With any luck it was a one off issue though it is worrying.

If you get the time, try an extended stress test of more than an hour.
Although running heaven for 45 minutes should have made the card crash if that was the issue.

At the end of the day if you don't feel comfortable with the card and have a bad feeling, return it.

I'm having a similar issue at the moment with my new vega 56 shutting my PC down. I replaced the
thermal paste on it last night, but that made no difference. Looks like i'll be returning a card

Hopefully you can get it all sorted out.

 

[Robbom]

Yeah I think I'm going to return the card, as I don't want issues down the line with no way to cover myself and not being able to find a solid fix hasn't really satisfied my worries.
I think i might just pay a bit more and get a new card, then I know that if I have the same issues not only am I covered for the card itself but I also can narrow it down to some other part or aspect of my pc
On that note the rx580 seems a bit better price/performance wise than what I have, but apparently has a higher power requirement, do you know whether my system (PSU especially) would support it or should I get another 1060?
Sorry to hear about your GPU, its frustrating when you just want hardware to work and it let's you down as I now know

 

[MetallicMonk]

Sounds like a good move. I would personally get another 1060 if I were you. I was originally going to get an RX 580 but went for the used vega instead
Your PSU is only 500W so would be at the edge of being able to run the 580. Would recommend upgrading it when you can.

Once you get your new GPU and give it a good test remember to come back here to see if that resolved the issue for people watching this thread.
I hope for your sake it does work.
If you could test the card in another PC or use another PSU that would narrow it down but it's not always an option.

Yeh hardware problems suck. My local PC store wanted to charge me $50 just to test my GPU...

Best of luck to you

 

[Robbom]

Yeah ok makes sense, managed to find a really good price on Amazon for a new card of the exact same model so that will really narrow the problem down.
Will definitely report back, thankful that this seller was a decent guy and has let me test the card before returning it as it has cleared up quite a few things.

Appreciate your help and hopefully I won't be coming back with more problems, good luck with your card to

 

[Robbom]

Well I've had the new GPU in for a day now, had the pc running for around 16 hours straight as I was downloading some large files on my slow internet and did some gaming and benchmarks last night that all went well so so far so good.
Something interesting I noticed when taking the card I previously returned out us that the board had some very small scratches on it that I hadn't previously seen and I'm assuming these could have signalled the problem. So I think the moral of the story is check used hardware thoroughly and if it seems too good to be true it probably is! Got a full refund no problems though
Just to piggyback off of this, I'm wondering whether to upgrade my CPU now as it's bottlenecking a bit on some more CPU intensive games.
My original idea was a 3770 non K which is fairly cheap used but I've now noticed that I could get a 3570k and overclock it for similar money or get a pretty much equivalent Xeon E3 1270 for around half as much. It seems the single core clock speed is only slightly lower than the i7 and it has a much better multithread score. I guess the question is what value should I put on 8 threads of Xeon and i7 over 4 threads on the i5 and is the Xeon is a good substitute for the i7 for the large gap in price/performance
Thanks

 

[MetallicMonk]

Hey there,

Really glad you solved the problem. The scratches might have been indicative of the card being mistreated so you've made a good decision.
As for your CPU. I can highly recommend the 3770 non K. I've been running this CPU for quite a while now and haven'r run into any bottlenecking.

Even with the vega 56. Which I am in the process of returning now also for a refund and replacement.

I would certainly put emphasis on 8 threads. More and more games are using multiple cores.
I'm not sure about the Xeon vs i7 argument. If i were you i'd do a bit of searching and see what other people think.

I use hwcompare and similar sites to compare processors and GPU. An i7 would be my recommendation, however I really don't
know enough about Xeons to definitively tell you which is better.

Just for interest, what score do you get on the 3d mark timespy demo benchmark? CPU and GPU

 

[MetallicMonk]

Oh and since your problem is solved you can mark the thread as answered
Hopefully the mods don't mind us using this thread off topic.

I would post the CPU question in the CPU section. There may be a lot more people to give you advice and help you make a better decision.

But I totally don't mind continuing on this thread.

All the best

 

[MetallicMonk]

my problem is SOLVED:

Was my brand new PSU. Using my old 550W until a replacement arrives. Did NOT think a brand new PSU would be responsible but it is!

Very glad it isn't the card or another system component.

 

[Robbom]

Ok cheers I will make a new thread on the xeon Vs i7 topic, thanks once again for your help and I'm glad youre getting your Vega 56 situation sorted out nicely! You got really unlucky with that PSU it seems
Oh and I'll grab that 3d mark result tonight

 

[MetallicMonk]

Ok cheers I will make a new thread on the xeon Vs i7 topic, thanks once again for your help and I'm glad youre getting your Vega 56 situation sorted out nicely! You got really unlucky with that PSU it seems
Oh and I'll grab that 3d mark result tonight

No problem at all I'm happy to help. Thanks for the for marking my post. Yeh I can't believe it was the new PSU. Getting a refund and ordering an EVGA G2 750W.