Thread: Real-World Benchmark

Since real-world benchmarks target specific hardware, there is a need for a series of different real-world benchmarks to test one specific component. This is because, different hardware react differently to various applications. Therefore a collection of benchmarks meant for a common purpose gives a fair evaluation of the hardware. Some of the most common ones are CineBench R10, video encoding, file compression, audio encoding, ray tracing, game benchmarks and file transfer. To give you a clear idea, we shall discuss the components that use these benchmarks.

Processors
For processors, the benchmarks we generally use are CineBench R10, audio and video encoding, file compression and ray tracing. While CineBench R10 and PovRay push the limits of a processor, they also use negligible resources from a graphics card.

Even video encoding requires number crunching and it calls upon the processor to comply with a task.

In this test we use part (1 minute) of a video clip and transcode it as per the X-264 format. Settings are tweaked to cause maximum stress on the hardware (depending on its capability of course).



Then there is file compression where we use WinRAR to compress a compilation of multiple files that measure 200 MB. Here, the key is to use the 'Best compression' method which further stresses a processor.

Hard drives
Benchmarks for hard drives evaluate capacity, speed and I"!olse emission. Although these parameters are crucial, when testing hard drives, their read/ write speed matters the most. It gives a user an idea as to how much time it takes to transfer data to a hard drive and copy from it, Let's say you want to copy 5 GB worth of music from one hard drive to another. If the target hard drive has a slow write speed, copying will delay even if your source hard drive features very good read speed and vice versa. This is called bottlenecking.

Therefore for testing a hard drive, we first create a RAM drive on a computer. Since RAM is the fastest, this method rules out bottle necking. Two sets of data are then copied on the RAM drive. The first is a 1 GB single file and the second comprises some random files that collectively measure up t9 1 GB. Each set is then copied individually from the RAM drive to the hard drive that's being tested. The logged time is the maximum write speed of that hard drive. This process is then reversed to get the read speed. We call this the file transfer test, and it is used across, laptops, netbooks, nettops, and external storage drives. Since MP3 players and PMPs have storage media, they are put to this test for evaluating speed as well.