Notes on Calculating Computer Performance

This report explains what it means to characterize the performance of a computer, and which methods are appropriate and inappropriate for the task. The most widely used metric is the performance on the SPEC benchmark suite of programs; currently, the results of running the SPEC benchmark suite are compiled into a single number using the geometric mean. The primary reason for using the geometric mean is that it preserves values across normalization, but unfortunately, it does not preserve total run time, which is probably the figure of greatest interest when performances are being compared. Cycles per Instruction (CPI) is another widely used metric, but this method is invalid, even if comparing machines with identical clock speeds. Comparing CPI values to judge performance falls prey to the same problems as averaging normalized values. In general, normalized values must not be averaged and instead of the geometric mean, either the harmonic or the arithmetic mean is the appropriate method for averaging a set running times. The arithmetic mean should be used to average times, and the harmonic mean should be used to average rates (1/time). A number of published SPECmarks are recomputed using these means to demonstrate the effect of choosing a favorable algorithm.