G.Skill 6GB PC3-12800 Triple Channel DDR3 Kit Review

System Benchmarks

When running the SuperPI 32MB benchmark, we are calculating Pi to 32 million digits and timing the process. Obviously more CPU power helps in this intense calculation, but the memory sub-system also plays an important role, as does the operating system.


We have always been quite fond of this application because it provides excellent insight as how to the processing and memory sub-systems are performing, and it is also perfect for demonstrating the slightest performance variances. With the processor locked to 3.2Ghz, we can clearly isolate how SuperPI scales with the increase in bandwidth. Evidently the more MB/s the merrier in SPi, as demonstrated by the massive 22 second gap between the CL7 and CL10. Clearly, if you want to post impressive SuperPI 32M results, memory frequency is your friend.


The latest iteration of the popular system benchmark is PCMark Vantage from the Futuremark crew. The PCMark series has always been a great way to either test specific areas of a system or to get a general overview of how your system is performing. For our results, we simply run the basic benchmark suite which consists of a wide range of tests involving all the sub-systems of the computer.


Much to our surprise, the Memory Suite suggested that performance gap between the CL7 and CL10 memory speeds was a microscopic 0.004%. We were quite puzzled by this result to be honest, and it seems like the Memory Suite is significantly CPU dependent. Since our Core i7 965 was clocked at 3.2Ghz for all four memory configurations, this hypothesis would explain the seemingly identical scores.


Developed by MAXON, creators of Cinema 4D, Cinebench 10 is designed using the popular Cinema software and created to compare system performance in 3D Animation and Photo applications. There are two parts to the test; the first stresses only the primary CPU or Core, the second, makes use of up to 16 CPUs/Cores. Both are done rendering a realistic photo while utilizing various CPU-intensive features such as reflection, ambient occlusion, area lights and procedural shaders.


Cinebench is a phenomenal CPU benchmark and the Intel Core i7 absolutely dominates in this test, but as this chart indicates the memory sub-system had truly minimal impact on the results.


Unlike many audio conversion programs, Lame Front End is not multi-threaded and only utilizes a single processor core. This will obviously limit performance but we should still achieve significant time savings going from the stock to the overclocked settings. We will be encoding a WAV rip of Santana’s Supernatural album and converting it to MP3 using the VBR 0 quality preset.


A second here, a second there, LFE does not appear to gain much of a performance boost from the increased bandwidth. We can't really complain though, since we are able to rip and convert an entire CD album to high-quality MP3s in under 2 minutes 40 seconds.



For the image editing portion of this review, we will use Photoshop CS3 in coordination with our very own HWC Speed Test, which is an excellent measure of CPU power and memory bandwidth. This is a scripted benchmark that individually applies 12 different actions and filters to a 3.72MB JPEG, and uses Photoshop’s built-in timing feature to provide a result at each test stage. Then it’s simply a matter of adding up the 12 results to reach a final figure.


Photoshop demonstrates some worthwhile performance scaling with the increase in memory bandwidth. A 7% gain with no increase in CPU clocks is definitely nothing to scoff at, and should convince image-editing enthusiasts not to short-change themselves with low-clocked memory kits.



The last of our real-life tests will be with the highly popular & multi-threaded WinRAR 3.71 tool, in which we take a 1GB batch of assorted files and archive them, timing the task until completion.


WinRAR compression is a heavily memory bandwidth bound workload, therefore it is no surprise that the DDR3-1866 CL10 setting outperforms the DDR3-1066 CL7 by a healthy 5% margin.