There is no standard computer.

The runtime measurement represents actual CPU time on whatever computer did the work. So, runtime is a measure of the effort put into computing the results, not a measure of the actual work represented by the tasks. We have points for that. Points should be proportional to the actual flop count of the task - the same points are awarded whether the task runs on a slow computer or a fast computer. But, the slow computer will report a higher runtime measurement (it takes longer to complete).

Hope that helps clarify things for you.