Now for the portion of the review you came for, performance. To reiterate from the Test Methodology and Specification, after the best orientation is determined we cover a minimum of 3 fan speeds per cooler using a common fan and at least one set of mounts when a fan is supplied in the retail package.
The Noctua NH-U12P comes with not only one fan, but two so we have single and dual fan tests including both the stock (NF-P12) and common fans (Scythe Gentle Typhoon 1850). Before we start covering the test run data, here are photos of the tested and best performing orientation along with a TIM print from the orientation testing.
If you are familiar with the EVGA X58 Classified, you know the orientation of the CPU socket, if not then this orientation is what I call “Fan parallel to socket arm” which is the best performing orientation for the Noctua on our bench and will be used for all logged test runs. You cannot tell in the photo, but the socket arm is just behind that last stick of RAM and under the fan mounted to the cooler. The TIM spread quite well for all mounts on the NH-U12P. As with all of our air cooler testing, we set aside the TIM that comes with the cooler and use Arctic Cooling MX-3 for all tests. For those of you who have worked with MX-3 before you know firsthand of the consistency which I can only describe as an oily, clay like combination. When I first opened a tube and started using MX-3, I thought maybe I got a bad batch… but no, each tube had the same consistency. For more info on Arctic Cooling MX-3, check out our TIM Round-Up V1.1.
Starting out, we mounted up a single NF-P12 at full RPM and kicked off our long test procedure. The NF-P12 is rated for 1300RPM at full voltage, and surprisingly our CFA-635 was able to run the Noctua fan at a 1394RPM average across all three mounts. I spoke on how simple the mounting system is on the NH-U12P, and achieving a 0.4C average core temp variance across all three mounts shows that simple also means effective for Noctua. The single stock fan brings forward a 67.56C average core temp for the compiled mounts, which was a little higher than I anticipated, but the 920 D0 I have on the bench is a toasty little chip.
Now to mount up the second NF-P12 for push/pull… Once again we maintained a tight mount variance of 0.6C, and stayed above the fan rating logging 1383RPM. Looking at the results, our core average only dropped 2C—from okay 1.89C if you are splitting hairs—with the NF-P12 in push/pull and maintain a near silent setup.
The stock NF-P12’s that come in the NH-U12P retail package are very capable fans and full out maintain silent to near silent operating noise, which is certainly a positive. Additionally, the NH-U12P completely stock was able to hold the heat load from my abnormally warm i7 920 D0. Time to set the stock fans aside and strap some GT-1850’s on…
With the stock fan data above, we can move on to the common fan for all coolers, the Scythe Gentle Typhoon 1850 (AP-15), this way we can provide a comparison of every cooler. Not to get ahead of ourselves, we need to cover the individual results before jumping to a comparison. So, on to the single GT-1850 at full RPM that does turn out a slightly better core temp average across the three mounts versus the dual stock fans. And once again, we hold a tight mount variance of 0.36C.
Turning the GT-1850 down to 1300RPM we see average core temps nearly match the NF-P12, with the stock fan having a slight advantage. As per the norm thus far, very little mount variation.
Cranking the GT-1850 as low as we go to 900RPM, we see the average core temps jump over the 70C threshold to 71.13C. Our mount variation gets a little squirrely as well, with 1.36C variance across the three mounts. I expect to see slightly higher variation as our temps increase, and 1.36C is still quite acceptable for mount variation.
For our next set, we mount up the second GT-1850 in pull to work with the GT-1850 already in push to see how well the NH-U12P can dissipate the heat on our toasty 920 D0. Not to sound like a broken record, but I expected to see a degree or two lower and I will cover my thoughts as to why this did not happen in the conclusion. However, the GT’s at full RPM do drop our temps over a degree lower than the NF-P12’s at full RPM which does fall in line with the single fan scaling. Mount variation swings to 0.9C which is definitely acceptable and continues to speak well for the mounting system.
Bringing the GT’s down to 1300RPM we get a little surprise with the stock NF-P12’s boasting a lower core average by 1.02C but I did was not expecting that outcome. Mount variation continues to stay solid with a 0.85C variance across the three mounts.
Slowing our Gentle Typhoon 1850’s down to 900RPM for both push and pull we stay under 70C for our core average with 69.71C. Mount variation held true, with 0.97C across all three mounts.
With the onslaught of individual results presented, time to bring the data into one chart and compare the Noctua NH-U12P to another cooler that can handle push/pull and this time we compare to a Prolimatech MegaShadow. For the comparison, we will include the single fan and dual fan data from both the Noctua and MegaShadow including the stock fan for the NH-U12P. Okay, time for the comparison chart…
Some of you may look at the chart and claim an unfair comparison since the MegaShadow is known for high end performance, but to be fair we used the same GT-1850’s for both coolers and present an apples to apples comparison. After studying the chart one thing should be clear, the Noctua is optimized for lower air flow where the MegaShadow performs better with more air flow as the core averages are quite close for single GT-1850 at 1300RPM.