PocketMatrix

The VOBenchmark reports the number of times a specific task can be completed in a set period of time. Hence, higher numbers denote faster performance.

Synthetic benchmarks are tests that measure some particular aspect of a system (memory bandwidth, CPU integer execution speed, etc.) rather than measuring how long a particular operation takes when running an application (file save, spreadsheet calc, etc.). Synthetic benchmarks are certainly worthwhile, but probably not so much to the end-user. The end-user cares about how fast something like PocketQuake will run on a new device, not how fast a CPU can read/write to system memory.

Most of the devs that hang out here can tell you that synthetic benchmark results are very meaningful (provided we wrote them). We can look at synthetic benchmark numbers and have pretty good idea of how fast our code should run because we know how much its performance depends on memory bandwith, CPU integer execution, cache misses, etc.