I received and have been using it stock, very happy. MSI Afterburner one-click OC Scanner (Hold Ctrl+F to access) gave me 2040-2085 MHZ (mostly 2070's) and 64c peak on GPU chip while playing BFV, at 100% power limit and +500 memory. I didn't try raising power limit beyond 100%, or memory above +500 yet. Stock fan settings. I had to stop using EVGA Precision X1 because it tried to use just one fan instead of the three, and thankfully I only let it stay that way for about 2 min before shutting down BFV. Resetting PX1 fixed it but still, scary since I didn't touch the fan settings.
 
I am going to test for longer than I originally planned because I am afraid of dead GPU after so and so days hearsay. After I test stability for a week or so, I will try doing the thermal paste and thermal pad changes, and put a 140mm fan on top of the back heatsink plate. Then I will report back here with thicknesses, dimensions, and differences in temps.
 
The cooler is larger than Arctic Accelero IV so I will save that for another GPU. 1080 Ti Kingpin copper heatsink is heavier but performs not as good as this FTW3 Ultra cooler. This cooler is fantastic and definitely keeping it on. My only complaint is that on my motherboard, the FTW3 Ultra doesn't snap into the PCB to trigger the locking mechanism due to how thick the cooler is. It barely fits in my quite decent size case, a testament about how far EVGA went to make the FTW3 Ultra cooler as big and low noise/temp ratio as possible. Only way I see Kingpin 2080 Ti's cooler being better is if they make the heatsink copper like they did for 1080 Ti, put in expensive thermal paste and pad.

memory temps?

scorpion73
memory temps?

 
Using the same settings as above including stock fans. I waited 10 minutes for temps to stabilize. This is what I got.
 
62c GPU middle
64c GPU bottom right
41c left memory
69c top memory
70c right memory
 
2070 MHZ average (MSI Afterburner OC scanner one-click)
About 270W usage
 
I am afraid of how fast this cooler cools down the card once GPU load stops. With default non-OC fan curve it just plummets from 64 to 50 in some seconds, and then to 45. That constant ramping up and down may not be the best for longevity.
 
Just for fun, I also ran a 55% GPU load program, and saw memory go to 70c top and 74c right memory. This is because GPU is cooler, fan 1 stops, and fans 2 and 3 spin slower accordingly. Memory still heats up at this time. So be careful with trying to go with super quiet build/no fan at certain low GPU temps.
 

if i set the fans to be always at 50% minimum without stopping them when GPU is cold at which temp the memory will work?

scorpion73
if i set the fans to be always at 50% minimum without stopping them when GPU is cold at which temp the memory will work?

depends on your case airflow and ambient temps. Check icx

I just modded with high end Noctua thermal paste and 11w/mk thermal pads. Don't do it. Temps are about the same, just 0-1c better at best.
 
What did you guys at EVGA use for paste and pads? Whatever the paste it's doing similar to Noctua, and I don't believe the pads are 1.2w/mk like I've read stock EVGA thermal pads are.
 
Here's dimensions:
 
Top of the PCB front plate with thermal pads touching GPU cooler:
7/8" by 2 3/4", 1mm (really weird. Falls through front plate to touch top of some chips)
1/2" by 4 1/2", 1mm (top of VRM sandwich)
 
Back of the PCB front plate:
3/8" by 4 1/2", 2mm (some small chips beside R22 chips)
9/16" by 2 1/4", 2mm (THREE of these for VRAM)
1/4" by 1/4", 2mm (one chip)
 
PCB:
5/16" by 4 1/8", 1mm (column of small chips)
1/2" by 4 1/8", 1mm (replaces the stock thermal paste above VRM)
 
Backplate:
7/8" by 4 1/4", 2mm (VRM)
3/8" by 1 7/8", 2mm (THREE of these for VRAM)
1 3/8" by 1 3/8", 2mm (GPU)
 
And again IMO it's definitely not worth switching out.
 

...be happy!!

https://youtu.be/dhmryCQAqiE

Did someone actually try the Thermalpad-Mod for the Memory Modules? I am kind of eager trying this, because I am seeing Temps well beyond 70+° C from 2 of 3 Banks (up and right). I will install some Thermalpads on the Baseplate, so that the Heatsinks actually touches the area over the Memory-Modules (like the Thermal-Mod from the EVGA GTX 1080/1070 FTW back in the days) and compare Temps before and after Benchmarking-/Gaming - scenarios. I wil let you know if that helps improving Temps :). I mean 74°C is still fine, but cooler is better! I noticed that the Fans are way quieter and more efficient than the ones they used on the ftw3 Version on the GTX 1080 Ti, hence I bumped up the Speed a bit. At first i was confused that they were starting to spin, then stop, then spin, then stop...etc until I saw that the Fan-Curve starts at minimum 20% (you can see the spotted line at 20%). The least you can set them to without being retarded is 18% which are roughly 540 RPM.
 
Cheers!

My memory temps may have gone down 1-4c depending on location, while GPU temp maybe 0-1c.
 
Make sure the new pads aren't thicker and to make sure the sandwich is pressed together tightly when tightening screws, or the GPU may have less than optimal pressure. I had that issue when I tried to redo my application of thermal pads+paste a few times with thicker pads to see if I did something wrong the first time around.

I experienced roughly the same. An additional Thermal Pad between Heatsink and Baseplate reduced the Memory Temp about 5 Degrees which is fairly OK I guess.