At this moment I'm trying to implement SIQS in Java in order to add

it to my factorization applet located at:

http://www.alpertron.com.ar/ECM.HTM

I finished the sieving part of the program. For the number 10^67+1 it

takes about 2 hours in a Celeron 333 MHz to find 5400 smooths (where

the upper bound is 118051). With this number of smooths the number

can be factored with the linear algebra phase, which should take only

a small fraction of this time (I haven't programmed it yet).

Of course, with modern microprocessors the sieving stage should be

much faster (possibly about 15 minutes).

Because of the memory limitations of Java I couldn't use the large

prime variation of the SIQS which should make the program run twice

as fast. I will need to use block Lanczos algorithm for the linear

algebra phase that is difficult to program but I have no other choice

because of the memory limitation.

For the numbers of the size noted above, if they have two factors of

about the same number of digits, the applet will factor tens of times

faster than now.

Best regards,

Dario Alejandro Alpern

Buenos Aires - Argentina

http://www.alpertron.com.ar/ENGLISH.HTM