- There are 3 major tests (registry, floating-point operations & integer operations).
- Each test is built based on some equations & operations depending on the test.
- Every major test contains 3 identical tests. For every identical test the seconds & milliseconds are calculated. At the end of a major test, an arithmetic average is made from the seconds & milliseconds times saved from those 3 tests. So if u get a wrong time at a "mini"-test the other 2 can make the error be ignored. Also the time is calculated with 0xx millisecond precision (at the start time must be x.0 seconds, so the last time cannot make a diference bigger than 0.9 seconds for tests). To be more exact the average from the milliseconds time is calculated. If the value is bigger than 500 ms it means test time has an error so it is corected adding an extra second to the test time.
- Score for the major tests is calculated with an simple equation (REVEALED NOW!!! 18.02.2005*).
- The final score = 40% Score Test1 + 80% Score Test2 + 80% Score Test3
- The program was compiled with no optimization (speed, floating-point), so tests cannot be in favour of a certain processor type.