Artifacts or voltage, what's worse? OC GTX 970

botchling86

Prominent
Jan 27, 2018
5
0
510
Hey! So Im new to overclocking but decided to see what I could do with my gtx970 and appears you can achieve a lot :)

HERES what I managed with for anyone interested:

MSI Afterburner Settings:
110 % power limit
+210 core clock
+500 Mem clock
+0 Voltage
Fan set to auto
Temp limit set to 85C


Ive used Valley benchmark and done some stress tests in MSI Kombustor and here are some GPU Z Readings after playing The Witcher 3 1440p DSR for a couple of hours:

Core clock max reading: 1552 mhz
(though decreased to 1539mhz in game for some reason, what does that mean?)
VDDC max: 1.2180 V
TDP max: 107.5 % (usually around 99-104 % in game)
max 67% fan speed
mem controller load max reading: 76 %
Max temp: 74C (side panel of the chassi thrown to the side. right now reciding on the extensive pile of laundry that someone just cant be bothered with)
FPS stayed around 40-45! :)


Has been almost completely stable without even touching voltage but.. just have a few questions. And they are most likely stupid and therefore maybe hard to answer... ready??

1. I see artifacts in the Witcher 3 (little red dots) they appear very seldom but now and again. Is this anything to worry about? I mean I have no crashing at all and havent touched voltage.

2. Whats worse for the graphics card.

-Artifacts?
-Extra voltage?
-Or is it mainly temperature?

3. Considering how well this card overlocks, about how far could I push the core clock to really? Maybe have to increase the power limit further for it to be any point? and is this even a good idea?

Thanks so much for reading. Would love an answer to this.
 


They all need to be avoided. What's worse? Having a cold or the flu? Extra voltage(.25mv increments) isn't necessarily bad for the card. There is a line where too much voltage can be very bad for the card. Overclock too far and you usually start with artifacts. They can get worse or the system may freeze or crash. The OC is removed and the OC'er tries again. It won't damage the card. Mess with the voltage without knowing what you're doing and you could physically damage a card.

GPUs usually run hotter than CPUs. You should always keep the card under 80C. The 70s are too warm for my tastes. The 50s and 60s(on air) is a GPU's Goldilocks zone imo. Now that I have a top end card with an integrated 120mm AIO cooler it generally stays in the 30's and 40's when gaming and around 25 at idle. With Nvidia's GPU boost 3.0 temperature is playing a bigger role than ever. It is similar to thermal throttling. That means that cooling is your only weapon against the clock changes.

I've heard of 58 and 62c. So now you're target temp should be 56.
 
Thanks for the quick reply!

When it comes to voltage, I thought the limit in msi afterburner (87mv) would be safe? so I just tried putting that on max right away to see if the artifacts would go away... no luck.

I was pretty careful at first looking for artifacts and ran both valley and some MSI tests and they were fine. I guess my question is If I should bother if they show up in some games. If there arent any crashes, should I still worry about damaging the card?
 


Can you take a screenshot of the snow? That would tell you if the artifacts are a cable error or a gfx card error. I had snow and it turned out to be a cable issue. If you see the artifacts and take a screenshot but it comes out clean that says replace your cable. If it turns out to be your GFX card I would start an RMA if possible. It has been a while since those cards were released but mine's still under warranty.
 
Thanks. I know its an OC problem.The artifacts (the red blinking dots) are known to appear when using overclocked maxwell gpus in the witcher 3. They ONLY appear in the witcher 3 and go away when I lower the clock. So RMA? No. Was just wondering if it was ok to ignore the artifacts since they dont appear too often. But I understand artifacts=uncommon stress on the gpu thus causing damage in the long run? So no option but to lower it I guess. 🙁
 


It does not mean uncommon stress. It means instability which imo would suggest that I talk to the manufacturer and ask for an RMA. Regardless, artifacts are never to be taken lightly or ignored or pushed aside.