This time I did the testing at 1300rpm. First I ran Valley with Kryonaut application from the last test and then switched to the Scythe paste. GPU-Z only showed a 1c increase in the average temperature but also a 10mhz increase in the average clock speed which is a bit odd.
Then I did another pass with the Scythe paste and switched back to the Kryonaut. This time I did 20 minute runs and GPU-Z showed a 2.5c decrease in the average temperature this time.
Screenshots here: http://imgur.com/a/bHjsx
So I don't know now. I think I likely just had a bad application of the Scythe paste in the first test as I don't think I've ever seen temperatures and fan speeds that high with Vsync on. That should have grabbed my attention in the first place.
------
Hi guys
I recently purchased the EVGA GTX 1060 SC and haven't been completely happy with temperatures and noise and have been experimenting a bit with it. I read some good things about Thermal Grizzly Kryonaut and saw it also had a few favorable reviews so I decided to give it a try. I was also curious as I never tried any alternative pastes before.
I did a little before and after test with it today and have the results to share. I searched the forum and haven't seen it mentioned much so I thought this would interest you.
So first, I'll go over my setup and testing procedure. I tried my best to be as methodical and consistent as possible. I wouldn't consider my test definitive though. I should repeat it at least three times but I don't have anymore Scythe paste and I also don't like taking apart my expensive GPU.
So I do recommend taking this with a bit of a grain of salt. I'm welcome to any suggestions or criticism here.First, I want to point out that this is a comparison to Scythe's paste that they include with their coolers, that comes in the little bag. I didn't notice any difference when I replaced the stock EVGA paste with it and in TomsHardware's Thermal Paste Comparison it did decent, 3c off the best performing paste. Link here: http://www.tomshardware.com/reviews/the ... 16-20.html
For the 3D application, I used Unigine Valley at 1440P, ultra settings, vsync on (60hz), and no MSAA. I chose this because in my experience it best represents a typical gaming session as this is the highest setting without dropping below 60 fps too often.
For the procedure, I ran Valley for 10 minutes, then opened GPU-Z and quickly alt-tabbed back to Valley and let it log for 15 minutes. Then I alt-tabbed to GPU-Z, changed the reading to average as quickly as a I could, and then took the screenshot and after that grabbed the maximum reading as well.
Here is a picture of the build that I tested it in: http://i.imgur.com/Kun8xhx.jpg The case is a Rosewill/Jonsobo Legacy U2. During the test the case fan was at 600rpm and the CPU at 800rpm. I do have the case side panel on.
For the application of the paste I use the rice method, without spreading it, and I use rubbing alcohol and coffee filters to clean the old paste. Here is a picture of the Kryonaut applied before fitting the heatsink back on: http://i.imgur.com/2yFtzbD.jpg
I changed the paste right after I finished the run with the Scythe paste and I started testing with the Kryonaut right after I installed it. It only took about ten minutes to change the paste and get the computer back on.
And lastly, the air temperature was 75F.
Okay, the results!
I've attached labeled GPU-Z screenshots below or you can view them on Imgur here: https://imgur.com/a/A9E9p
Between the average measurements there was a 2.7C drop, 200 RPM drop in fan speed, and a 4Mhz increase in clock speed because of GPU Boost 3.0. For the maximum measurement there was also a 3C and 200 RPM drop.
After I did this I wish I would've compared them at the same fan speed but what I did was set the fan speed to run static at the average speed with the Scythe paste, which was 1530 RPM, and ran Valley again. Here there was a 5.6C drop in the average temperature and a 20Mhz increase in the average clock speed. There was also a 5C drop and a 25Mhz increase for the maximum reading.
I think you could safely call that a 5C improvement. I'm pretty impressed by this and glad my money wasn't wasted. I paid $10.50 on eBay for the 1.0 gram tube which looks to be enough for 4-5 applications. I think $10 and very little effort and time for a 5c drop is pretty damn good and significant for the noise conscious, small form factors, and overclockers who are on the edge with temperatures or noise, which was my case.
And one last thing I want to mention; it might not be as effective on CPUs as most modern CPUs use an integrated heat spreaders where GPUs are a bare die - and I think CPU coolers typically have higher mounting pressures so the TIM might not make as much of a difference.
Well I think that covers everything. I hope you find this test useful and I'm curious to hear your thoughts about this or any other experiences with Kryonaut.
