Making strided arrays as fast as native arrays

The author fails to understand two critical factors:

1. Compilers have been applying the sort of "strength reduction" optimization he talks about automatically for about 20 years.

2. The performance penalty of strided memory access is due to cache issues, not due to address computation.