post #1 of 1
Thread Starter 
Hey friends,

I have a Galax GTX 1070 Mini I bought with the intention of using it in a small form factor build. My idea behind choosing the card (which might have been ill-advised) was that I could tune its performance and power consumption down--if I needed to--and probably still end up with a card more powerful than a 1060. Plus, it being a 1070, I figured I could always use it in SLI later on if that seemed like a good idea.

Anyways, the card's more powerful than my 980 like I was expecting it to. It runs Just Cause 3 in 1440p maxed out no problem, which is the most demanding game I own at the moment. Overall I'm happy with the card's out-of-the-box performance. Recently, though, I started trying to tune the card to learn about overclocking/underclocking and undervolting, and hopefully gain some insight into how much performance is traded off in pursuit of lowering power consumption.

(Here's a reddit thread I made over the weekend that touches on some of the problems I'm going to mention here.)

So, the main complication I ran into while I was playing around in MSI Afterburner last weekend is that my card will lock itself to any voltage and clock speed I set it to in MSI Afterburner until I get to 0.962V. It seems like any clock speed I set the card to at 0.962V* and lock it in, I have this problem. The card jumps down in voltage and clock speed pretty wildly, this causes the card to perform significantly worse than at lower voltages, where it remains at whatever clock speed and voltage I set it to.

I have a google spreadsheet where I was recording my data, and you can see I stopped recording at 0.962V, as from that point on my card's clock speed and voltage started fluctuating.

It was suggested to me that it could be that my card's temperature reaches a certain temperature and starts to throttle itself via GPU Boost, but I don't think that's the case. I wasn't recording temperatures, but I'm pretty sure that my temperatures were getting higher earlier in my testing when I hadn't yet locked my fan to 85% (when I noticed my temperatures were getting into the 60Cs), and since I set my fans to 85%, I don't think the card's gone over ~55C.

Speaking of the fans on my GPU, is there any way I could tune them to spin up/curve according to something other than the card's temperature? If I could have the card's fans spin according to power usage or clock speed I feel like the card could much more responsively and appropriately keep itself cool.

And the last question related to my GPU and its fans, the card doesn't allow the fans to go below 30%. I'd like for them to be able to spin down further or turn off when the card is idle. I emailed Galax asking if they could update their BIOS and they pretty much said "No, that's a feature of our EXOC SNPR cards." Galax's 1070 EXOC SNPR is, of course, also a 1070, and it also has two fans, but it has different video I/O. Will I mess my card up if I flashed to that card's BIOS? Are there any other BIOS I could get away with? (This is something I'm not worried about fooling with until I tune the card "before" so I can compare the before/after!)