Test Systems

In order to test PCMark 10, three TechPowerUp editors worked together to put together a total of four different hardware configurations ranging from a Mini PC to full-blown gaming PCs with Intel, AMD, and NVIDIA hardware alike. Thanks a lot to crazyeyesreaper and cadaveca for their help here. The four systems used are specified below:

All four systems were kept at stock settings and were not overclocked for maximum stability and to ensure replicability. With the advent of the GPU Boost technology from NVIDIA, actual boost clocks may well differ from advertised boost clocks, so please be aware of that. PCMark 10 as well as PCMark 10 Extended benchmarks were run on all four systems as we felt including PCMark 10 Express was redundant. Each system was checked to have all critical updates for the OS installed. All the latest hardware drivers were also installed, other non-essential running programs were closed, and each benchmark was run multiple times. As we found out, PCMark 10 is very reproducible. Multiple runs differed by ~3-4% on average, and as we will see on the next page, a representative score for each system was reported where a lot of tests are common to both benchmarks and where our test systems were thus scored twice each.