Closed lducas closed 3 years ago
So far, we don't do any update_gso
calls for safety in any of the BKZ routines for performance reasons, so I'm not sure this is the right thing™ to do. At least we should measure the performance impact.
Ain't there a mechanism to recompute only what is actually not up to date ?
Le mar. 12 janv. 2021 à 10:40, Martin R. Albrecht notifications@github.com a écrit :
So far, we don't do any update_gso calls for safety in any of the BKZ routines for performance reasons, so I'm not sure this is the right thing™ to do. At least we should measure the performance impact.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/fplll/fpylll/pull/204#issuecomment-758533918, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABQGTYFFROM33EJIB4XGYHLSZQKKVANCNFSM4V6YZ6AA .
n_known_rows
is a thing so this might be the thing to branch on, it's protected though? We'd need to expose it to FPyLLL/Python first, though.
Ok, nvm.
A tentative fix for issue #203 .