The test rig consists of the following parts:
Processor: Intel Core i7 3770K
Motherboard: Gigabyte G1.Sniper3
Video Card: EVGA GTX 690
Memory: Samsung Green MV-3V4G3D/US – 16GB
Power Supply: Enermax Maxrevo 1350W
Storage: OCZ Vertex 3 240 GB raid 0 Running Windows 7 64bit
Cooling: Deepcool Maelstrom 240
Case: Phanteks Enthoo Luxe
As always testing was done using Intel Burn Test (IBT) with the AVX instruction set. The reason to use this instead of Prime95 with this setup is that it pushes our i7 3770K a whole lot harder. This gives us a better sense of extreme load temps. Most of the testing was done at the max memory setting for a single pass. While admittedly this in no way would qualify the chip as stable while overclocked, it gives us a realistic idea of what temps it will be hitting. Additionally we are going to run 3DMARK Fire Strike Extreme to give you a realistic idea of performance while gaming.
Our baseline was established using a bequiet! Dark Rock Pro 2 with the 3770K on stock settings:
Ambient temps: 22.22C/72F
Idle temps: 38.75C/101.75F core average
IBT temps one pass max memory: 63.5C/146.3F core average
Since the Kraken X41 is designed to be used with the CAM software we used it for all of our testing. It is a very well designed suite packed with real time graphs and all of the software controls. You can adjust the fan speed and have full control of the pump LED from within. I especially like the CPU core temperature reporting that automatically averages your core temps for you. This is great when you just want to take a quick look at your overall temp. In addition to all this, NZXT has also developed a mobile app that allows you to access CAM without having your mouse in hand. For our purposes I set our profile to “performance” but still used Hardware monitor to gather our temperature data and got to work.
And here are the results using stock settings and the Kraken X41:
Ambient temps: 19.44C/67F
Idle temps: 33.5C/92.3F core average
IBT temps one pass max memory: 57C/134.6F core average
At this point I’m liking what I’m seeing. Adjusting for ambient, we still have and idle temperature 4.45° F cooler than our baseline as well as a solid 6.7° F cooler during IBT. Next test please!
As opposed to simply going for a max overclock, we are simply going to shoot to hit 4.5ghz pass or fail. If the cooler will allow our 3770K to hit 4.5 ghz we note the temperatures and move on to test it on 3DMARK Firestrike Extreme. If not then back down to a more acceptable and stable overclock. Not as extreme, but slightly more consistent and far more analogous to normal use.
Here are our results for our i7 3770K @ 4.5ghz with 1.248V vCore:
Ambient temps: 19.44C/67F
IBT temps: 75C/167F with one core hitting 79C
We are very pleased with the numbers put down by the X41. While these are not our very best over all when adjusted for ambient, they are by far our best for a single fan unit. It looks like that extra radiator thickness really does go a long way. And with our max core temp of only 79 we have extra headroom for a higher overclock if we decide we need one.
With a passing result from IBT, we went ahead with a single run through 3DMARK Fire Strike Extreme @4.5ghz:
Ambient temps: 19.44C/67F
3DMARK: 62.5C/144.5F with two cores hitting 64C
So there you have it, the X41 will crank through your favorite game with the skill and ease of a seasoned veteran.
In addition to performance we always have to consider what is the cost. That cost for coolers is usually noise. Is the noise of the unit acceptable/bearable for the level of performance you achieve? The X41 is a mixed bag in this category. We messed around with the profiles for the fans preloaded into CAM, but found that only performance mode really gave us the cooling we were looking for. But with this comes the typical whooshing sound that 140mm fans make when at speed. We do prefer this to the high pitched whines from smaller diameter units as it more easily becomes background noise, but it’s still there. Overall I give it a solid B for noise and can live with it for the results.
And what about that claim of being the worlds first AIO with a variable speed pump? Well in CAM you will see a designator for PUMP RPM. While testing this number would fluctuate up and down with the temperature, so we know that it does in fact work. What we didn’t notice was whether it really mattered. At no point did we ever notice that the pump was overly loud. In fact at no point did we every really notice pump noise. So it either works great, or it’s just a really quiet pump to begin with.
Finally we have the LED on the pump. This is pretty neat (although not a first) in that it can be changed to any color on in the spectrum as well as pulse and change colors in a cycle. All this is controlled through CAM of course. An odd thing we found though is that since the LED is powered (seemingly) through the USB header it never turns off when connected to our test motherboard. This motherboard has a feature that detects power draw through USB and allows devices to charge even when the machine is turned off. As a result the LED stays lit. Not a problem for us but maybe something to consider?