[Cado-nfs-discuss] A problem on polySelect

Zimmermann Paul Paul.Zimmermann at loria.fr
Sat Jun 25 09:34:14 CEST 2011


       Dear Esmaeil,

> Dear Friends
> 
> I want to factor this C100 number
> "1323457225332313736475970722598968952464464764081931733627176279763066404693603990852604111129229583"..
> but i got a warning that is "DEAD" on polyselect phaze.
> 
> i found the file "c100.kjout.0-20000" on tmpdir directory.  [file
> attached ( <1kb )]
> 
> I used released and svn version on this number but it is same.
> 
> Is this a bug? Please guide me.
> 
> Thanks.  Esmaeil

did you succeed in factoring the 59-digit number in README?

What are your hardware, compiler, operating system?

What is the exact command you gave with the svn version? On my laptop (64-bit
Core 2) I get with svn revision 856:

[zimmerma at coing cado-nfs]$ ./build/coing.loria.fr/polyselect/polyselect2 -q -kmax 20 -incr 60 -admin 0 -admax 20000 -degree 4 -maxnorm 36.8 -t 1 -N 1323457225332313736475970722598968952464464764081931733627176279763066404693603990852604111129229583 50000
# ./build/coing.loria.fr/polyselect/polyselect2 -q -kmax 20 -incr 60 -admin 0 -admax 20000 -degree 4 -maxnorm 36.8 -t 1 -N 1323457225332313736475970722598968952464464764081931733627176279763066404693603990852604111129229583 50000
# Initializing primes took 2ms
# Tried 333 ad-value(s), found 50 polynomial(s), 14 below maxnorm
# potential collisions=6.77e+01 (3.78e-01/s)
# best logmu: 33.19 35.01 35.17 36.08 36.19 36.31 36.41 36.47 36.50 36.58
# Total phase took 179.14s
n: 1323457225332313736475970722598968952464464764081931733627176279763066404693603990852604111129229583
skew: 196032.000
# lognorm: 33.32, alpha: -2.71 E=30.61
# Murphy's E(Bf=10000000,Bg=5000000,area=1.00e+16)=9.91e-09
c4: 11400
c3: 597361891
c2: -26406808357172
c1: -301224112598576316060
c0: -5716667522987085405047409
Y1: 842342791253
Y0: -583715541055761932510492
m: 397472778221307430608889748235599716008144253913395814396288561718836644161229443109481252934523651
# generated by ./build/coing.loria.fr/polyselect/polyselect2.r846 -q -kmax 20 -incr 60 -admin 0 -admax 20000 -degree 4 -maxnorm 36.8 -t 1 -N 1323457225332313736475970722598968952464464764081931733627176279763066404693603990852604111129229583 50000 in 179.15s

To help isolating the problem, you might add the following lines just before
the "assert (u < ppl)" line (773) in polyselect/polyselect2.c:

             if (u >= ppl)
                  printf ("p=%lu ppl=%lu rj=%lu rq[i]=%lu u=%lu\n",
                          p, ppl, rj, rq[i], u);

Then do "make" again, run the above polyselect2 command, and send us the
output you get.

Paul Zimmermann



More information about the Cado-nfs-discuss mailing list