Our testing methods
Obsessive readers may note that the components used in this build differ slightly from those that Cyril uses in his case reviews. The table below shows the specifics of the hardware we'll be using for our test bed.
|Processor||AMD Phenom II X4 965 Black Edition (140W)|
|Processor cooler||Thermaltake Frio (single fan in a pull configuration)|
|Memory size||4GB (4 DIMMs)|
|Memory type||Corsair XMS3 at 1333MHz|
with default Windows drivers
|Graphics||XFX Radeon HD 6870 1GB GDDR5
with Catalyst 11.5 drivers
|Hard drive||Seagate NL35.2 500GB 7,200 RPM|
|Optical drive||Asus DRW-1814|
|Power supply||OCZ GameXStream 700W|
|OS||Microsoft Windows 7 Professional 64-bit|
In an attempt to promote some consistency, the chosen parts for this system use roughly the same amount of power as Cyril's at full load. Using a Kill-a-watt P3 meter, I measured the following peak power utilization numbers (at the wall) to use as a reference.
|CPU load only||302W|
|GPU load only||280W|
|CPU & GPU loads||394W|
Due to the similar energy usage, you can compare these test results to Cyril's with the requisite salt shaker in hand. To make things scientific, however, I'm maintaining a separate data set going forward, representing only the cases I've tested using these parts in the same environmental conditions. The components used may not be the newest kids on the block, but they have approximately the same power and thermal characteristics as today's high-end hardware.
Most of the tests and methods employed are publicly available and reproducible. If you have any questions about our methods, hit up our forums to talk with us about them. Below is a list of the relevant software pieces used in this review.
|Fractal Design's Define S is a leaner, meaner R5||28|
|Android Wear gets always-on apps, Wi-Fi support||4|
|Tiny MIT touchpad fits on your fingernail||16|
|OnePlus One sales now open to everyone||22|
|AMD says Windows 10 will be released in late July||29|
|Happy 50th birthday, Moore's Law||46|
|This week produced a bumper crop of security holes, patches||17|