I think my 570 is on the way out.

daphatgrant

Fi-yah!
Staff member
Joined
Jun 15, 2003
Messages
18,564
While browsing the web today the cards driver apparently stopped working. I later tried to play a couple games and got terrible artifacting and then freezing. I have tried reinstalling one of the games, updating the cards drivers with no changes. I just ran OCCT 4.0.0 here is what I ran:

V-Card%20Issue.jpg


After 1 min I had over 3700 errors.



So... I'm assuming it's time to box this bad boy up? It's an EVGA superclocked GTX570.
 
I had an EVGA GTX570 that went out on me after 2 weeks. First video card I ever had die on me.

EDIT.. I know this doesn't help your situation but man it pissed me off...
 
Last edited:
you might need to tune the voltage.

On the card itself or in OCCT?

What really throws me off is that the card has been working fine until this morning.

As of right now nothing I attempt to play will run for long before artifacting and eventually freezing.
 
Get MSI afterburner, and see your voltage and tell us what it is.

but was this problem from the day you got it or did it just start happening now out of the blue?
 
Your VRMs may have gone bad. Seems to affect a number of 570s FWIR, though usually only when the cards are voltage tuned or aggressively clocked.
 
Did you re-seat the card and re-seat the power to the card? Don't see your setup in your sig.. any watercooling? If so check for leaks etc. Or try in another machine if you can.
 
I just sent my EVGA 570 back last week after it failed exactly like you described above after 2 months of use. Randomly my video drivers quit working then artifacts in the millions.

I tried old, new, beta etc. drivers, re-seating, different mobo etc. still same results.
 
Yup, sounds like your card shit the bed. I would just run a few runs of Linpack (Linx/IBT) just to be sure that the card is at fault.
 
I measured my GPU for almost 45 minutes to try and weed out an odd issue. My VIN6 power rating continued to drop the entire duration of the benchmark. What are the VIN1, VIN4, and VIN6 measurements?
 
Well, I called EVGA and talked to them for a bit, they had me restart and go into the bios so I could read off the voltages to them. I had restarted previously but still had the issues, when I went to restart for the EVGA tech I accidentally shut the computer off, after finishing the with the tech I finished the post from a complete shutdown and everything seems to be working. If all I needed was a shutdown and not a restart to fix the problem I'm a bit embarrassed.
 
That's odd that a restart would fix it. I wouldn't expect video card errors to be fixed with a shutdown.
 
Well, I called EVGA and talked to them for a bit, they had me restart and go into the bios so I could read off the voltages to them. I had restarted previously but still had the issues, when I went to restart for the EVGA tech I accidentally shut the computer off, after finishing the with the tech I finished the post from a complete shutdown and everything seems to be working. If all I needed was a shutdown and not a restart to fix the problem I'm a bit embarrassed.

Its a very common problem with GTX570 and it shouldn't be happening. If any one should be embarrassed its Nvidia.
 
Glad to hear a simple restart was a quick fix for you, Ive had other odd issues with my 570 that required a full power off on cycle to come back to right myself. I was playing MW3 (I know Im lame dont bother saying so) and the card refused to clock up past the 2d core clock of 401 and restarting drivers and Afterburner had no effect at all...frustrating niggles I know.

One thing Im noticing from the picture is your card is a non-reference pcb and Ive heard of people having odd issues with these cards while looking for voltage to clock references in my overclocking adventures.

These VRM failure stories have me scared as well, I got the cheapest 570 with a lifetime warranty I could find so I know mine has no upgrades to the VRM.
 
Glad to hear a simple restart was a quick fix for you, Ive had other odd issues with my 570 that required a full power off on cycle to come back to right myself. I was playing MW3 (I know Im lame dont bother saying so) and the card refused to clock up past the 2d core clock of 401 and restarting drivers and Afterburner had no effect at all...frustrating niggles I know.

You obviously didnt read this thread at all, because a shutdown was what fixed his issue, not a simple restart. This is what threw him for a loop
 
Thank you Mr. troll cirial...but obviously you didn't read my post?

"Ive had other odd issues with my 570 that required a full power off on cycle to come back to right myself"

A comma is not a period.
 
Its a very common problem with GTX570 and it shouldn't be happening. If any one should be embarrassed its Nvidia.

Agreed. I periodically have the same problem, however as of now, it does seem a shutdown fixes the issue.

Glad to see this thread though. I was afraid maybe I had a bad mobo.
 
The vrms and overall failure rate of the 570s is pretty bad. That is why I went with a 480gtx soc. Get a card later in the cycle on sale with aftermarket cooling is my thing.

To the op. Your card is going to die. No way a restart is anything more than a temp fix.
 
67918638.jpg

62767237.jpg


This was how I found my EVGA 570 card yesterday. The sticker on the cover fell off. Luckily I haven't experienced any errors yet.
 
EVGA is really good with RMAs at least. I have the same card as you, got it two days ago. Hopefully i don't have the same fate but if I do at least I know EVGA service is great

actually i have the non hd sc
 
Yeah, I've read about GTX500 cards that develop artifacting that survives across reboots, but is cured with a quick power-down. Sounds like the card's memory controller might have some power-on-only sticky state that gets corrupted, but who knows.
 
I had a video card with heatpipes and no fan. It was supposed to be a passive 3D card but when I put a fan on it, the pink spots I saw on the desktop dissapeared. So, maybe heat might be a factor.
 
Do you keep your computer on 24/7?

If I keep my computer on for too long my card will start acting funny. I just shut down for the night every couple days and my card will work perfectly. Otherwise, it'll have crazy artifacts in the web browser, games will crash, my computer will randomly reboot, etc.

A quick shut down may or may not work. I've tried just shutting down then booting up a couple minutes later and typically the problem is still there. You may have to keep it off for a while.
 
Well here's a little update, the card has been running great since that last set of issues until now, for some reason it now crashes the whole computer when I play Dead Island. The timing of the crashes is pretty random ranging from 2 hrs to 5 mins of playtime. I installed EVGA's Precision monitor and the whole time I played Dead Island my temps stayed right around 85C. To test the card I played Crysis 2 for 2 1/2hrs and again the temps never exceeded 85C (Crysis 2 never crashed). Right now I am running the EVGA OCC scanner for 20min and again the temps have never exceeded 85C and there are no artifacts recorded. On Mon I think I'm going to call EVGA and see if I can't upgrade to a 580. I've had multiple EVGA cards 8800gts, GTX 285, and this is the first problem I've had. If this is a GTX 570 issue I'd like to get away from them, hopefully a 580 will perform correctly.
 
85C??? thats pretty darn hot.

you have your fan running on max speed on that thing?

i'm not saying that running that hot will cause crashes but if your card is starting to get flaky getting it to run cooler could possibly fix it. personally i hate how hot my 560 Ti's run when on auto fan speed because i know long term heat on any PC component will result in failure so i keep the fans locked to 100% so they never go much higher than 65C under full load.
 
85C??? thats pretty darn hot.

you have your fan running on max speed on that thing?

i'm not saying that running that hot will cause crashes but if your card is starting to get flaky getting it to run cooler could possibly fix it. personally i hate how hot my 560 Ti's run when on auto fan speed because i know long term heat on any PC component will result in failure so i keep the fans locked to 100% so they never go much higher than 65C under full load.

85'C is well within spec for the 570. Auto fan speed won't even start really ramping up until 88-89'C.
 
85'C is well within spec for the 570. Auto fan speed won't even start really ramping up until 88-89'C.

yeah i know this i'm just saying i won't ever let my video cards get that hot on a regular basis, its simply not good for them IMHO

also auto fan speed ramping up at that high of a temp is utterly pointless in my experience, once a card gets that hot the only way to cool them down is to shut down the load that made them get that hot, set fans speeds to max and then try again. i experienced this with my 8800 GT with its stock cooler some games took it to 100C, if i set the fan speed to max before hand rather than letting the card do its thing it would keep it from ever getting near that extreme temperature. i eventually bought an aftermarket cooler for that card because it ran so damn hot and guess what it lasted a very long time and afaik the guy i sold it to a year ago is still using it without problems.
 
If there is a particular fan speed that can keep the temps below 70c then there is no need to go 100% speed, you will have to test it. I use that graph thingy where you plot what fan speed should be when temps reach a certain degree. but i dont think heat is your problem, 85c is okay, sort of. You havent downclocked the voltage and other stuff? Thats the first thing you should try.
 
Back
Top