non monotonic B1. Why?
In commit f0b52916 (in 2013), a sequence of ECM curves with increasing B1 was introduced.
However, that came after a curve with a larger B1, so we have non-monotonic B1s.
This behaviour is still present with the current version, in sieve/ecm/facul.cpp
. It is visible with -v -v
in las, e.g. with:
$build_tree/sieve/makefb -side 1 -lim 1200000 -poly ./parameters/polynomials/F9.poly -out /tmp/F9.roots
$build_tree/sieve/las -lambda0 2.2 -lambda1 2.9 -I 12 -poly ./parameters/polynomials/F9.poly -lim0 2300000 -lim1 1200000 -lpb0 26 -lpb1 24 -mfb0 52 -mfb1 72 -fb1 /tmp/F9.roots -q0 1200000 -v -v -nq 1
Here are the lines of interes.
# make plan for P-1 with B1=315, B2=2205
# make plan for P+1 with B1=525, B2=3255
# make plan for ECM with B1=105, B2=3255, parameterization = 8, parameter = 1, extra primes = 1
# make plan for ECM with B1=315, B2=5355, parameterization = 8, parameter = 2, extra primes = 1
# make plan for ECM with B1=115, B2=5775, parameterization = 8, parameter = 3, extra primes = 1
# make plan for ECM with B1=125, B2=6195, parameterization = 8, parameter = 4, extra primes = 1
Simple question: why? (@zimmerma ?)