News for hackers. Stuff that matters.
Re:The Factoring Dead: Preparing for the Cryptopocalypse by poulson on Sunday August 04, 2013 @06:33PM (#6157366)
Interesting. That isn't excessively large for distributed-memory Krylov-subspace eigensolvers (e.g., SLEPc). Do you have a good reference for the current algorithms used? And how large is the nullspace typically?

EDIT: I assume this [1] is what you're referring to, which draws from the following algorithms of Coppersmith [2,3].

[1]: http://eprint.iacr.org/2010/006.pdf

[2]: https://www.sciencedirect.com/science/article/pii/0024379593...

[3]: http://thales.doa.fmph.uniba.sk/macaj/skola/seminar/Coppersm...

Re:The Factoring Dead: Preparing for the Cryptopocalypse by pbsd on Sunday August 04, 2013 @08:32PM (#6157744)
Yes, that is correct. Generally the block Lanczos variant used in practice is Montgomery's [1], though, not Coppersmith's.

Lanczos was popular until a few years ago, as it requires slightly less actual matrix-vector multiplications, but since Wiedemann can be partially distributed it is becoming more desirable.

[1] http://link.springer.com/content/pdf/10.1007%2F3-540-49264-X...