04-14-2020, 01:00 AM

On the website

https://www.thimet.de/CalcCollection/Cal...mance.html

there is a benchmark written in a number of languages and run on a good selection of calculators. One problem, though, is the computation of the Performance Index, P.

The site shows this:

P = 34 / T

where T is the time the benchmark took. The performance indices shown, however, use this formula:

P = 34 * N / T

where N is the number of loops. In most cases, these will be the same since N = 1 but quite a few of the examples use more than one loop. That must be taken into account to get the Performance Indices shown in the timing tables.

https://www.thimet.de/CalcCollection/Cal...mance.html

there is a benchmark written in a number of languages and run on a good selection of calculators. One problem, though, is the computation of the Performance Index, P.

The site shows this:

P = 34 / T

where T is the time the benchmark took. The performance indices shown, however, use this formula:

P = 34 * N / T

where N is the number of loops. In most cases, these will be the same since N = 1 but quite a few of the examples use more than one loop. That must be taken into account to get the Performance Indices shown in the timing tables.