You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think a challenge with these parameters is that for "typical" choices of, say "beta" and "d" these aren't even uSVP instances, i.e. the norm of the target will not be unusually short.
Yeah, good point, these are definitely edge-case params.
In terms of a fix, other than 8cd0c51, the only thing that comes to mind is that we could somehow bound σ/q -- and not give certain attack outputs (e.g. BDD/uSVP) when the bound is satisfied?
Dual/mitm/exhaustive-search should still be able to provide some sort-of useful output, right?(unless I'm missing something).
Running:
Gives an output of:
bdd
seems very low here, a few things seem odd:β: 40
η: 258
which is greater thann=256
)I guess (1) and (2) are probably related.
The text was updated successfully, but these errors were encountered: