980 ti woes

agnes · 11 Feb 2016 at 22:29

I do get this VERY consistent artifact in Tomb Raider, actually stopped playing it because of it... Wonder if it's the cause lol.

I've uninstalled Tomb Raider for now, as nothing I did fixed it and it was doing my head in...

Might reinstall and try again after increasing volts? It's a seriously long shot. But who knows.

Gregster · 11 Feb 2016 at 22:47

If increasing the volts keeps it stable, that is your answer and I personally would return the GPU as it isn't able to hold the boost clocks on stock voltage like it should.

agnes · 11 Feb 2016 at 22:48

Gregster said:
If increasing the volts keeps it stable, that is your answer and I personally would return the GPU as it isn't able to hold the boost clocks on stock voltage like it should.

If it is indeed the case, and tomorrow Heaven is still running fine. Then I'll return it on Monday (weekend yo... Gaming time!).

I don't mind being without it, however, up to 28 days :S

My life is about to get VERY productive.

Gregster · 11 Feb 2016 at 23:29

agnes said:
If it is indeed the case, and tomorrow Heaven is still running fine. Then I'll return it on Monday (weekend yo... Gaming time!).

I don't mind being without it, however, up to 28 days :S

My life is about to get VERY productive.

Yer, the dreaded RMA and if it helps, I had an RMA take 6 weeks with Asus

tommybhoy · 11 Feb 2016 at 23:45

Thought I read somewhere that Evga does an advanced swap out(don't know what it's actually called) where they advance charge your cc and send out a new gpu and they kept the gubbed one, obviously you'll need to ask them first.

agnes · 12 Feb 2016 at 07:54

Well, with the +18mV on. The test ran fine. Until 0330 when my PC decided to reboot for updates... Super annoying.

But yea, seems it's the GPU. I'll send it off next Monday

Danny75 · 12 Feb 2016 at 08:10

Don't rule out the memory altogether. I've got RAM that passes Memtest and stress tests no problem at its rated 2400Mz frequency, which started causing all sorts of errors and crashes after two years and I (found I) had to lower it to 2133MHz which sorted everything out (was driving me nuts for weeks). With another batch of RAM in another build, it would also pass Memtest without errors but was the cause of the crashes in that case, and swapping it out for a new (identical brand/model) pair fixed it.

Hope you find the fix, whatever it is.

Toss3 · 12 Feb 2016 at 08:18

agnes said:
Well, with the +18mV on. The test ran fine. Until 0330 when my PC decided to reboot for updates... Super annoying.

But yea, seems it's the GPU. I'll send it off next Monday

Did you try it without extra volts? Seems like a weird issue to have because of a faulty GPU. Seems more like a ram-issue, or just regular instability. Try setting every voltage in you bios manually.

agnes · 12 Feb 2016 at 09:17

Toss3 said:
Did you try it without extra volts? Seems like a weird issue to have because of a faulty GPU. Seems more like a ram-issue, or just regular instability. Try setting every voltage in you bios manually.

I did try without the increased mV and within 20 mins, got the .exe has stopped working error.

I'll triple check tonight to make sure I can fully replicate. Then, if I can, I'll put the GPU into my girlfriends machine, run the EXACT same tests. If it's the same thing. Then it's 100% my GPU.

Spikey · 12 Feb 2016 at 09:21

Definitely seems like a GPU issue to me, RMA time!

Kirby Wurm · 12 Feb 2016 at 11:03

So it's the voltage? I've also had that same issue with a factory OC Gigabyte 560 Ti I borrowed from a friend while my GTX 570 was being RMA'd a few yers ago. With a small voltage bump in Afterburner it was stable.

Smffy · 12 Feb 2016 at 11:23

agnes said:
If it is indeed the case, and tomorrow Heaven is still running fine. Then I'll return it on Monday (weekend yo... Gaming time!).

I don't mind being without it, however, up to 28 days :S

My life is about to get VERY productive.

Advanced RMA on a 980Ti is €750 deposit, i'm doing one now... Expect it refunded in 1-3 days from when you return the original faulty card. Have you tried a different slot on the PSU out of interest - just in case there is some form of voltage variance in it before you RMA the card.

Glad you found the issue though!

agnes · 12 Feb 2016 at 13:12

Smffy said:
Advanced RMA on a 980Ti is €750 deposit, i'm doing one now... Expect it refunded in 1-3 days from when you return the original faulty card. Have you tried a different slot on the PSU out of interest - just in case there is some form of voltage variance in it before you RMA the card.

Glad you found the issue though!

DAMN fine suggestion on the different slot on the PSU. I'll try that tonight.

As for the advanced RMA. Thanks for the heads up.

Chris

Kirby Wurm · 12 Feb 2016 at 16:13

Glad you're nearing a resolution. The troubleshooting process can be a painstaking nightmare.

agnes · 12 Feb 2016 at 16:17

Kirby Wurm said:
Glad you're nearing a resolution. The troubleshooting process can be a painstaking nightmare.

I'm a senior techie in IT, I've grown to love and hate the troubleshooting side of life... After a decade in this career though, it can be a bit monotonous.

Midlife Crisis · 12 Feb 2016 at 16:50

I have a GTX 275 golden sample that would only be stable if the memory was down clocked by 15mhz, more power made it worse. I should have RMA'ed that but I put up with it

Drollic · 12 Feb 2016 at 22:48

agnes said:
I'm a senior techie in IT, I've grown to love and hate the troubleshooting side of life... After a decade in this career though, it can be a bit monotonous.

I thought about selling my current rig due to it. All because my board doesn't like 1 specific LLC setting when overclocking - setting 2 of 5. It took hours to find as the cpu was stable under load, prime95 etc. I've also had the above with a 7950 on this same rig, unstable at default memory clock, took a while to pinpoint.

Lokken86 · 12 Feb 2016 at 23:10

Intermittent problems like that are a pain. I suspect if it isn't the psu I'd be looking at the motherboard, they can often die slow deaths and give problems like your describing.

Dirk Diggler · 12 Feb 2016 at 23:48

I've had that "has stopped working" issue with tomb raider a couple of times. It happened with my MSI 6G and also with my Classy. I've now got the Classy 24/7 running +100 core and +300 mem. This wasn't stable initially, but I've slid the voltage slider in AB right up, and also the temp/power target right up.

It's been solid 1518 on the core with those settings and no crash yet. I've only had the Classy about a week but so far so good.

I think it's caused by the auto throttle the cards do due to temperature or power limits. However, with my MSI I ran it stock and it never crashed on that setting. If the Classy starts playing up, I'll put that back to stock as the OC is pure vanity really.

To summarise: try raising your temp/power target and also up your volts.

Drollic · 13 Feb 2016 at 00:01

String said:
I've had that "has stopped working" issue with tomb raider a couple of times. It happened with my MSI 6G and also with my Classy. I've now got the Classy 24/7 running +100 core and +300 mem. This wasn't stable initially, but I've slid the voltage slider in AB right up, and also the temp/power target right up.

It's been solid 1518 on the core with those settings and no crash yet. I've only had the Classy about a week but so far so good.

I think it's caused by the auto throttle the cards do due to temperature or power limits. However, with my MSI I ran it stock and it never crashed on that setting. If the Classy starts playing up, I'll put that back to stock as the OC is pure vanity really.

To summarise: try raising your temp/power target and also up your volts.

Sounds like the all or nothing approach to reach 1500+. Wouldn't you be better off settling for 1450-1480 ensuring peace of mind? :eek: