This number could certainly be tested, but before someone jump up and
starts testing it using the current PFGW, please speak up that you
want to test it. Because of the form, pfgw would currently use the
"generic" reduction code. However, with some modifications to the
code, the GFN DWT code can be made to process this number, by simply
working in 2^8388608+1 "space", and then performing a "final" modulus
to (2^8388608+1)/(5*2^25+1) to get the final result. Using this
code (the DWT) will net a ~6x improvement in speed.
My guess is that the "generic" reduction code would run in 6 months
to a year, while the DWT would take a month or two of runtime.
It would not take too long to write code to perform:
and get the full benefit of the DWT. NOTE that the only thing that
can be determined is the F23 factor is composite, or probably prime.
I don't think there is any way to know if this number is prime. I
do believe this is just a little out of range for Primo ;) and don't
know if there is anything like Pepin's test for factors like this.
--- In primenumbers@y..., Pavlos N <pavlos199@y...> wrote:
> Hello all,
> I was wondering if the number (2^8388608+1)/(5*2^25+1)
> can be tested with openpfgw.The character of this
> number(composite or prime) is not known,i
> believe.Perhaps there is someone brave (having spare
> time and cycles of course)enough to test it