I have just returned from vacation and noticed the discussion about the
problems with sieving on a P4.
The P4 is an excellent sieving platform with a proper SSE2 implementation.
Paul Jobling and I have agreed to implement SSE2 into the sieving core of
NewPGen when he comes back from vacation. It may take me some time to
develop the SSE2 core for modes other than k.2^n+/-1 so be patient.
For those of you who are considering a computer purchase and wondering how
the P4 will perform in sieving, I have temporarily make my personal sieve
(TwinGen) with the SSE2 implementation available. It only has k.2^n+/-1
modes implemented. You can access it at
When NewPGen is updated to SSE2, it should be faster than TwinGen, and
should have all its operating modes. TwinGen is available to give you a
lower limit idea of the speed of a P4 in sieving.
I have been an Athlon supporter for 2 years now. I still like Athlons.
This implementation along with the SSE2 implementations for PFGW and Proth
make the P4 a very attractive prime computer.