[Retros] Natch Euclide etc.

[Marco Bonavoglia]

> For the record: Natch on P4 3 GHz 512 MB

> A: 1 mn 18.13 s

> B: 10 mn 11.87 s

> not a great diff

> /anders uddgren


With a AMD64 (3200+) is more or less the same (slightly better) BUT if you
don't use the hash-tables it improves dramatically

A: 14.01 seconds
B: 4 mn 10.93

how's with a P4 ?

It seems that for the above problems the overhead of using the hash table is
predominant, or possibly the new processors are too fast and hash tables
aren't necessary?
Is the above already known? May be Natch could be better parameterized?

Marco Bonavoglia, Milano, Italy