Re: Not good for Ada endorsement

[Lucretia <laguest9000@googlemail.com>]

On Friday, 9 July 2021 at 10:21:05 UTC+1, Jeffrey R. Carter wrote:
> On 7/9/21 11:16 AM, Jeffrey R. Carter wrote: 
> >> if Number mod 2 /= 0 and then Sieve.Bits(Index_Type(Number)) then 
> >> 
> >> and then incrementing Number by 1. 
> > 
> > Since both operands are positive, mod and rem give the same results, and rem is 
> > usually a bit faster. It's normally not an issue, but in this case it's done 
> > billions of times, so it might make a difference.
> Note also that short-circuit forms (and then) require disabling processor-level 
> optimizations and may have a negative effect on execution time when used 
> unnecessarily.

You don't need those as the algorithm always starts at 3 and increments by 2.

I managed to get just under 4000 passes with a 1 bit array, but not using Ada's packed array. That's actually the slowest method, strangely.