Is the GTX 1080ti watercooling a good investment?
Is the GTX 1080ti watercooling a good investment?
I discovered during my research on overclocking that I adjusted certain settings for my graphics card. I couldn't push it beyond +75MHz, though the improvement was noticeable compared to other models. By modifying the graph to allow higher voltages past 2000MHz, I managed to increase its speed further. However, once it hits 55*C, it only stabilizes around 2025MHz. There were still other voltage settings available in the data, but I didn’t explore them because performance was already limited by temperature constraints. In reviews about overclocking with a watercooled 1080Ti, they often mention MHz increases using Afterburner without changing the voltage options in the graph.
The graph I mentioned is what I was discussing. It seems the 1.093v lock-off might be present, though it appears to stabilize at that point, possibly explaining why. Initially, I was restricted to the +75MHz O/C range, but I shifted the settings later, which helped push the clocks higher by tweaking MHz/voltage on the chart. I’m not sure if this is unique, but in all my research on overclocking these cards, the only reference I found was a single Reddit discussion. Any articles or videos I’ve seen about watercooling and overclocking are mostly using Afterburner sliders rather than adjusting voltage stages in that specific graph. Since I managed to boost clocks by modifying the graph, I’m curious if others with active cooling setups can do the same.
I’m also looking into an actively cooled configuration, not just standard waterblock and radiator, but one that would work well with SLI.
Biglizard, I haven’t checked CPU-Z to determine what might be limiting my card. I understand that NVidia throttles these cards between 55*C-60*C, so they don’t get as hot as the Maxwell chips. I’m aware I’m experiencing thermal throttling, so I haven’t spent much time pushing overclock limits without trying to lower temperatures more. My goal is to find out how far back the card sits from the back plate, and I’d like to add a thermal pad between it and the back plate, along with a large heat sink for the back plate and smaller ones for the VRMs—I think those are the main heat sources.
It really comes down to whether you have the funds for a full build. If so, you won’t need to invest more money, but it’s not worth the effort, cost, or minor performance gains. I personally don’t upgrade my GPU unless I really need it. Running just a few games is enough, and a small difference in frames won’t make much of a difference unless you’re consistently hitting 30-40 FPS. Save your money.
Don't waste time cooling it down. Modern graphics cards handle temperatures well, and you won't lose much performance from this. Put the money into a better monitor or SSD instead. This claim is completely wrong. With Pascal, temperatures matter a lot. The lower they are, the better the boosts will be, and the card will run more smoothly.
If you wish to reduce the temperatures of the card, feel free to discuss it over a break (via a full cover block and a custom loop or an AIO cooler with a bracket), but keep in mind that the card was designed by NVidia and has its own threshold. Only certain cards, like the K|NGP|N Edition from EVGA and possibly the HOF cards from Galax, are expected to perform better when overclocked. This information isn’t reliable. The main difference between these GPUs and others is their increased power phases and slightly higher power limits, which won’t make a significant impact unless paired with advanced cooling solutions.
With a 1080 Ti, you can either use a Shunt mod or flash the XOC bios to bypass the power limit. But because NVIDIA still restricts voltage to 1.093v, you won’t reach around 2100 without special modifications and significant changes. A stock FE card will give you about 2000mhz.
In spite of all that you've said, Vellinious, the truth of the matter is that no matter how cool you keep it(without using exotic methods of cooling) there is yet a ceiling implemented by Nvidia to safeguard the company and/or the end user. That limit is what you're seeing all across
this thread
. In order to break those limits you're in essence voiding your warranty when you make any form of physical modifications to the GPU.
Aggressively overclocking their components, unless stated by the OP that they are expendable, isn't a worth while suggestion to an end user.
Lutfij :
In spite of all that you've said, Vellinious, the truth of the matter is that no matter how cool you keep it(without using exotic methods of cooling) there is yet a ceiling implemented by Nvidia to safeguard the company and/or the end user. That limit is what you're seeing all across
this thread
. In order to break those limits you're in essence voiding your warranty when you make any form of physical modifications to the GPU.
Aggressively overclocking their components, unless stated by the OP that they are expendable, isn't a worth while suggestion to an end user.
Actually, no. That's again, inaccurate.....the cooler you keep it, the less voltage you'll need for a specific clock. By lowering the temps, and thus lowering the voltage, you can avoid the early power limit hit. You'll still eventually hit it on reference boards, but with most of the custom PCB GPUs, they have higher power limits set from the factory....they won't hit them as soon. I've had my 1080s running as high as 2278 and didn't hit a power limit and I don't hardware mod.