User talk:Jzakiya/sandbox

Hi, please excuse my intrusion here on your page, if you think it's inappropriate just say so. First, try not to pay attention to the unfortunate rudeness of some WP users. Just ignore that. Second, if I may, some crucial data is missing in your article. There is no analysis for the complexity of your algorithm. At the very least, maybe you could run it few times for varying problem sizes like 10^7, 2*10^7, 4*10^7, 8*10^7, so that the empirical local complexity could be measured, just to get a rough idea of your algorithm's performance, as log(t2/t1) / log(n2/n1) ( log(t2/t1) / log(2),   in the above cases)?

As for primality testing, doesn't your algo still need all primes below the square root potentially? Doesn't it test the divisibility against each of them? Doesn't it make it a kind of trial division algorithm?.. How does it speed compare with Miller-Rain test, for some really big number, like 11^123+13^132, or some other tests? (try it e.g. in this page).

Cheers, -- WillNess (talk) 21:32, 11 December 2011 (UTC)