Discussion related to cado-nfs

[paul zimmermann <Paul.Zimmermann@inria.fr>]

       Hi,

> Hi, I am interested in evaluating the performance of personally-generated
> polynomial in real NFS.
> I've tested the CADO-NFS software and it works out great. But, I somehow 
> want
> to use it for polynomial testing.
> I have generated them (the polynomials) somehow using the development of
> Montgomery's method and want to test it out.
> Is there some way to test them using CADO-NFS? However, testing using CADO
> will be starting from sieving instead of poly. selection (because the poly,
> is obtained already)

I'm not sure what you mean by "test it out" but there are 2 ways you can test
polynomials you have generated:

1) you can compute their Murphy-E score and compare it with the one of
   polynomials generated by CADO-NFS (or other polynomials), for example:

$ ./build/tomate/polyselect/rotate -Bf 30940618 -Bg 17246818 -I 14 -skew 
10800 parameters/polynomials/rsa155.poly 0 0

n: 
10941738641570527421809707322040357612003732945449205990913842131476349984288934784717997257891267332497625752899781833797076537244027146743531593354333897
Y1: 1
Y0: -39123079721168000771313449081
c5: 119377138320
c4: -80168937284997582
c3: -66269852234118574445
c2: 11816848430079521880356852
c1: 7459661580071786443919743056
c0: -40679843542362159361913708405064
skew: 10800.000
# lognorm: 50.44, alpha: -6.23 (proj: -2.36), E: 44.21, nr: 5
# MurphyE(Bf=3.094e+07,Bg=1.725e+07,area=4.153e+15)=5.28e-11
# generated by 3d09b51cc: 10800 parameters/polynomials/rsa155.poly 0 0 in 
0.01s

  The higher the Murphy-E score is (here 5.28e-11), the better the polynomial
  is. Make sure to use the same Bf/Bg/I parameters when comparing two
  polynomial pairs, and to provide a correct skewness (if no skewness is given
  in the file, it will be automatically computed by CADO-NFS).

  However, this method does only work for polynomial pairs with the same
  degrees.

2) you can perform a sieving test with CADO-NFS, for example by importing the
   polynomial you have generated. See scripts/cadofactor/README. Then the ETA
   printed by CADO-NFS will indicate how good is your polynomial.

Best regards,
Paul Zimmermann