New
#421
P.s that is so funny, me and the girlfriend went to a local supermarket earlier and saw this old scraggly guy walking past us and after I said "did you see Willie Nelson walking past us?"
P.s that is so funny, me and the girlfriend went to a local supermarket earlier and saw this old scraggly guy walking past us and after I said "did you see Willie Nelson walking past us?"
Those days are over now though with what Nvidia has done. Unless they go back to the way things used to be, those overclocking times are over.
The classy voltage tool works with the 1080 Classy, and will work with the TI KPE. Overclocking isn't dying....it's just changing. Now, it's not about just throwing a metric ton of voltage at something, moving all the sliders to the right and praying. Now it's about monitoring temperatures, voltage, clocks and how each one of those things effect the other, in order to get the best clock possible.
I think Pascal is the most challenging, and frustrating GPU I've ever overclocked...but also very rewarding when done properly.
Voltage doesn't help....lower temps help. Without lowering temps to below 0c, adding voltage may increase clocks, but at the expense of added heat, which will lower effectiveness.
I've had 3 GPUs I ran with cold coolant, and all of them would run 2250+ on stock voltage....but at normal room temperatures, they'd only do 2202, sometimes 2214, but...never ran well there. They ran better at 2189 at normal room temps.
It's all about the temps.
I agree with you statement as many have said here before. I also believe you can get a good balance between Voltage and temp to achieve good clocks that can be sustained for the likes of benchmarking. Ideally the best scenario would be to give the GPU the maximum amount of voltage it can handle with the highest thermal limits allowed while having the actual temps as low as possible (Using some exotic cooler to bring the temps down to remove the heat issues) Just my thoughts.
I agree, for the most part. None of the pascal GPUs I've tested benefited very much at all from increasing the voltage above 1.093v until the coolant was running below 0c, and the core temps were kept at, or below 15c during the benchmark run. Sure, above that you could see some increased clocks, but the scores didn't scale with it...marginal improvements, at best. Really, anything above 20c on the core, you're just wasting time with added voltage.