Page history

rosettacode>Mwn3d

Somewhere in there I got rid of the need for 12 as a BigInteger, reformat

rosettacode>Mwn3d

Forgot to update the output, remove runtime check

rosettacode>Mwn3d

With the GCD check we can be sure that we have found a primitive, so we can mark them (negligible speed change)

rosettacode>Mwn3d

It turns out that the GCD check can speed it up a little bit

rosettacode>Mwn3d

This is what I was really looking for when I made that last optimization

rosettacode>Mwn3d

Another slight speedup

rosettacode>Mwn3d

Don't have to scale by 1. Negligible effect on the run time.

rosettacode>Mwn3d

Save another operation (this way might do the same thing under the hood)

rosettacode>Mwn3d

Possibly saving one addition every once in a while

rosettacode>Mwn3d

When I said "very very slight speedup" I was only looking at the perimeter < 1000 results. It actually makes a huge difference.

rosettacode>Mwn3d

a*b = 0 (mod 12) for all triples, very very slight speedup, checking that a*b*c = 0 (mod 60) actually slowed it down

rosettacode>Mwn3d

Moved the LIMIT declaration but forgot to take the comment with it

rosettacode>Mwn3d

Fursther optimization by skipping more a,b pair candidates (noted in the comments), rm runtime calculations

rosettacode>Mwn3d

Incredibly slight optimization (took the runtime of a 10k perimeter limit from 90 seconds to 89 seconds on this computer)

rosettacode>Mwn3d

Extra notes and comments, move a comparison so it is only calculated if needed.

rosettacode>Mwn3d

Notes about the Triple class, also THREE and FOUR were never used

rosettacode>Mwn3d

A version that uses brute force to find primitives