On 2009-Apr-07 21:49:12 -0700, William Stein <wst...@gmail.com> wrote: > >On Tue, Apr 7, 2009 at 9:24 PM, Bill Hart <goodwillh...@googlemail.com> wrote: >> The reason it runs slow is a.factor() is bizarrely slow in Sage. It's >> like a factor of 50 times slower than Pari. > >There are some differences: > (1) pari's factor is *not* provably correct, but Sage's is. [That >said, this will make no difference for small input.]
How often is pari's factor wrong and how? If pari's factor is mostly correct bu occasionally reports incorrect factors, given the massive discrepancy in speed, if it worthwhile always trying pari's fator first and then verifying its output. If it's wrong, then try Sage's factor. The only issue I can see would be if pari sometimes incorrectly reports composite numbers as prime - which means every "prime" number reported by pari would need to be fed to Sage. -- Peter Jeremy
pgpugCtZueiOZ.pgp
Description: PGP signature
