Test Setup

As you can see, the fan contraption barely fits on our testing station. In fact due to the one sided clips on the P55 board from Asus, there is nothing for the metal to hold unto, so it would just fall off in a normal usage scenario. Once turned on, the blue LEDs of the fan turn on as well. From an aesthetics point of view, the mix with red on the Trident modules and blue color LEDs in the fans do not make any sense. It would have been much cooler for G.Skill to use red LEDs instead. That said, the color happens to go well with the theme of the Asus P55 board in our case.

Performance & Overclocking

First off, this is a so-called high-capacity kit. This means that each module is 4 GB large and as we all know the bigger the modules the thougher they become to overclock. The same goes for the number of DIMMs in a kit. That said, getting the G.Skill Trident 8 GB kit to run at 2000 MHz CL9 is very easy by simply setting the XMP profile in the BIOS. There seems to be some headroom for overclocking, as I managed to hit 2133 MHz with them at this setting and 1.7 V. This seemed to be the end of the line, as higher voltage did not yield more performance with the system becoming unstable with just one or two MHz more on the base clock. The next step meant trying out CL5-5-5-15 to see if the kit could handle that. It was no big surprise that the system did not boot with that latency no matter what voltage was supplied. The first bootable scenario was a CL6-6-6-18 setting in combination with 1.6 V. Keeping the latency, each bump in voltage meant an increase of 20 to 30 MHz up to a maximum of 1474 MHz.

The step up to 7-7-7-21 bore the first pleasant surprise. The 8 GB kit booted right up at 1.5 V and 1600 MHz with no problems, but at such low voltage, 1660 MHz was the end of the line. Pushing it to 1.6 V also allowed us to run it at a slightly higher MHz, while the memory hit a ceiling at 1.8 V and 1712 MHz. Increasing the voltage did not result in better overclockability. Relaxing the timings to CL8-8-8-24 and climbing the voltage ladder the 8GB kit, I was suprised that it managed 1938 MHz. It would have been nice if the kit could have broken the 2000 MHz barrier at CL8, but no matter how much voltage we applied, the memory did not perform any better. Due to the high frequency, the multipier was changed from 2:10 to 2:12 during the benchmarking process at this latency. With CL9 the kit managed to push the limits of the entire test bench and ran fine at 2133 Mhz, but anything above that resulted in an unstable system. This is most likely due to the system itself and not the memory kit.




As you can see, the memory scales very well with voltage, climbing steadily when increased. The memory performance did not degrade past 1.7 or 1.8 V, but did not translate into higher overclockability.