Loading ...
Sorry, an error occurred while loading the content.
 

Wasn't this too slow?

Expand Messages
  • rstanile
    Hi, please take a short look at this (completed) factorization. Wasn t it too slow (on this fast machine)? Were the parameters correct? Should the matrix be so
    Message 1 of 10 , Aug 1, 2005
      Hi,

      please take a short look at this (completed) factorization. Wasn't it
      too slow (on this fast machine)? Were the parameters correct? Should
      the matrix be so big?

      Washuu

      PS. I tried ECM, I haven't just been lucky enough.


      N=184882746155407225169603381797719531303157417581
      05700458865665051663583049065157772996647203603442
      175230555693510905800477
      ( 122 digits)
      Divisors found:
      r1=3414377722997660306778420701787237313 (pp37)
      r2=5414829909125842652937999139221690431040170057
      783327360590432991821745452530835636829
      (pp85)
      Version: GGNFS-0.77.1
      Total time: 137.10 hours.
      Scaled time: 69.24 units (timescale=0.505).
      Factorization parameters were as follows:
      name: hp782
      n:
      18488274615540722516960338179771953130315741758105
      70045886566505166358304906515777299664720360344217
      5230555693510905800477
      skew: 97259.91
      # norm 2.22e+16
      c5: 3720
      c4: -4421372558
      c3: -133549990222290
      c2: 6810756205034300238
      c1: 1039633026003540222243229
      c0: 16169508505016419402458994044
      # alpha -5.06
      Y1: 3479345214763
      Y0: -346155890495632062217315
      # Murphy_E 2.22e-10
      # M
      70058916964555953120503678457052770951121009888205
      99316670227513859859737087520281507959970424794977
      230625625928268216714
      type: gnfs
      rlim: 5000000
      alim: 5000000
      lpbr: 27
      lpba: 27
      mfbr: 50
      mfba: 50
      rlambda: 2.4
      alambda: 2.4
      qintsize: 60000
      Factor base limits: 5000000/5000000
      Large primes per side: 3
      Large prime bits: 27/27
      Max factor residue bits: 50/50
      Sieved algebraic special-q in [2500000, 6400001)
      Primes: RFBsize:348513, AFBsize:348365, largePrimes:7890324
      encountered
      Relations: rels:8169525, finalFF:805105
      Max relations in full relation-set: 28
      Initial matrix: 696954 x 805105 with sparse part having weight
      83420732.
      Pruned matrix : 655747 x 659295 with weight 57255737.
      Polynomial selection time: 1.51 hours.
      Total sieving time: 121.22 hours.
      Total relation processing time: 0.72 hours.
      Matrix solve time: 13.10 hours.
      Time per square root: 0.54 hours.
      Prototype def-par.txt line would be:
      gnfs,121,5,maxs1,maxskew,goodScore,efrac,j0,j1,eSt
      epSize,maxTime,5000000,5000000,27,27,50,50,2.4,2.4,60000
      total time: 137.10 hours.
      --------- CPU info (if available) ----------
      CPU: Intel(R) Pentium(R) 4 CPU 3.40GHz stepping 04
      Memory: 1031516k/1048016k available (2084k kernel code, 15572k
      reserved, 769k data, 232k init, 130512k highmem)
      Calibrating delay loop... 6717.44 BogoMIPS (lpj=3358720)
    • trilliwig
      Hi Washuu. 137 hours for a C122 general NFS factorization sounds about right. I have done a C126 in 227 hours and a C121 using the older polyselect in 202
      Message 2 of 10 , Aug 1, 2005
        Hi Washuu. 137 hours for a C122 general NFS factorization sounds
        about right. I have done a C126 in 227 hours and a C121 using the
        older polyselect in 202 hours.

        The newest versions of GGNFS have updated the polynomial selection
        parameters so that it will take more time looking for a polynomial
        (about 7 hours instead of 1.5 for a C122). It's reasonable to think
        that the extra time investment will pay off by finding a polynomial
        good enough to reduce sieving time by a larger amount. The matrix
        perhaps could be made smaller at the expense of more sieving time, but
        from the looks of your statistics this is not likely to decrease your
        overall time.

        Now if your number was of special form, you could create a polynomial
        with much smaller coefficients and the factorization would be a lot
        faster. A number of SNFS difficulty 122 would take about as much time
        as a GNFS factorization of an 80-digit number. But only if your
        number was of special form.

        --
        Sam

        --- In ggnfs@yahoogroups.com, "rstanile" <washuu@e...> wrote:
        > Hi,
        >
        > please take a short look at this (completed) factorization. Wasn't it
        > too slow (on this fast machine)? Were the parameters correct? Should
        > the matrix be so big?
        >
        > Washuu
        >
        > PS. I tried ECM, I haven't just been lucky enough.
        >
        >
        > N=184882746155407225169603381797719531303157417581
        > 05700458865665051663583049065157772996647203603442
        > 175230555693510905800477
        > ( 122 digits)
        > Divisors found:
        > r1=3414377722997660306778420701787237313 (pp37)
        > r2=5414829909125842652937999139221690431040170057
        > 783327360590432991821745452530835636829
        > (pp85)
        > Version: GGNFS-0.77.1
        > Total time: 137.10 hours.
      • Geoffrey Reynolds
        If your P4 has hyperthreading (don t all the new ones have hyperthreading?), enable it and run a seperate sieving process on each virtual CPU. This should
        Message 3 of 10 , Aug 2, 2005
          If your P4 has hyperthreading (don't all the new ones
          have hyperthreading?), enable it and run a seperate
          sieving process on each virtual CPU. This should
          reduce sieving time by a third or more.

          --- rstanile <washuu@...> wrote:

          > --------- CPU info (if available) ----------
          > CPU: Intel(R) Pentium(R) 4 CPU 3.40GHz stepping 04
          > Memory: 1031516k/1048016k available (2084k kernel
          > code, 15572k
          > reserved, 769k data, 232k init, 130512k highmem)
          > Calibrating delay loop... 6717.44 BogoMIPS
          > (lpj=3358720)


          Send instant messages to your online friends http://au.messenger.yahoo.com
        • Bouk de Water
          Smart idea! It works here for me. Question: q0 sets the starting point for seiving. Is there a possibilty to set an endpoint to avoid double work?? Bouk. ...
          Message 4 of 10 , Aug 3, 2005
            Smart idea! It works here for me.

            Question:

            q0 sets the starting point for seiving. Is there a possibilty to set an
            endpoint to avoid double work??

            Bouk.

            --- Geoffrey Reynolds <g_w_reynolds@...> wrote:

            > If your P4 has hyperthreading (don't all the new ones
            > have hyperthreading?), enable it and run a seperate
            > sieving process on each virtual CPU. This should
            > reduce sieving time by a third or more.
            >
            > --- rstanile <washuu@...> wrote:
            >
            > > --------- CPU info (if available) ----------
            > > CPU: Intel(R) Pentium(R) 4 CPU 3.40GHz stepping 04
            > > Memory: 1031516k/1048016k available (2084k kernel
            > > code, 15572k
            > > reserved, 769k data, 232k init, 130512k highmem)
            > > Calibrating delay loop... 6717.44 BogoMIPS
            > > (lpj=3358720)
            >
            >
            > Send instant messages to your online friends http://au.messenger.yahoo.com
            >
            >
            >
            > Yahoo! Groups Links
            >
            >
            >
            >
            >
            >
            >


            __________________________________________________
            Do You Yahoo!?
            Tired of spam? Yahoo! Mail has the best spam protection around
            http://mail.yahoo.com
          • trilliwig
            ... There s no option right now to do that... though support for a qend could be added to the script without too much trouble. Usually the multi-client mode
            Message 5 of 10 , Aug 4, 2005
              --- In ggnfs@yahoogroups.com, Bouk de Water <bdewater@y...> wrote:
              > Question:
              >
              > q0 sets the starting point for seiving. Is there a possibilty to set an
              > endpoint to avoid double work??

              There's no option right now to do that... though support for a qend
              could be added to the script without too much trouble. Usually the
              multi-client mode of the script works sufficiently well that it's not
              needed, though I realize this requires a shared directory to work.

              --
              Sam

              > --- Geoffrey Reynolds <g_w_reynolds@y...> wrote:
              >
              > > If your P4 has hyperthreading (don't all the new ones
              > > have hyperthreading?), enable it and run a seperate
              > > sieving process on each virtual CPU. This should
              > > reduce sieving time by a third or more.
              > >
            • Bouk de Water
              Hello everybody! I have a question. Is it possible to use a second irreducible polynomial with ggnfs? If so, how do I enter it in the *.poly file? Say I wanted
              Message 6 of 10 , Aug 12, 2005
                Hello everybody!

                I have a question.

                Is it possible to use a second irreducible polynomial with ggnfs? If so, how do
                I enter it in the *.poly file?

                Say I wanted to use these two polynomials for a Lucas number:

                f(X) = 6*X^6 + 27*X^5 + 45*X^4 + 15*X^3 + 30*X^2 - 3*X + 4
                g(X) = 12200160415121876738*X - 19740274219868223167

                with:

                m=128812498419496899286734041447037910680019851818697424172503181354467240508369208475713248057682151827591523707537;

                How do I proceed? (if possible)

                Bouk de Water.


                __________________________________________________
                Do You Yahoo!?
                Tired of spam? Yahoo! Mail has the best spam protection around
                http://mail.yahoo.com
              • Bouk de Water
                Answering my own question perhaps :-) It s Y0 and Y1 isn t it, the non-monic parameters? Bouk. ...
                Message 7 of 10 , Aug 12, 2005
                  Answering my own question perhaps :-)

                  It's Y0 and Y1 isn't it, the non-monic parameters?

                  Bouk.

                  --- Bouk de Water <bdewater@...> wrote:

                  >
                  > Hello everybody!
                  >
                  > I have a question.
                  >
                  > Is it possible to use a second irreducible polynomial with ggnfs? If so, how
                  > do
                  > I enter it in the *.poly file?
                  >
                  > Say I wanted to use these two polynomials for a Lucas number:
                  >
                  > f(X) = 6*X^6 + 27*X^5 + 45*X^4 + 15*X^3 + 30*X^2 - 3*X + 4
                  > g(X) = 12200160415121876738*X - 19740274219868223167
                  >
                  > with:
                  >
                  >
                  m=128812498419496899286734041447037910680019851818697424172503181354467240508369208475713248057682151827591523707537;
                  >
                  > How do I proceed? (if possible)
                  >
                  > Bouk de Water.
                  >
                  >
                  > __________________________________________________
                  > Do You Yahoo!?
                  > Tired of spam? Yahoo! Mail has the best spam protection around
                  > http://mail.yahoo.com
                  >
                  >
                  >
                  >
                  > Yahoo! Groups Links
                  >
                  >
                  >
                  >
                  >
                  >
                  >


                  __________________________________________________
                  Do You Yahoo!?
                  Tired of spam? Yahoo! Mail has the best spam protection around
                  http://mail.yahoo.com
                • gchil0
                  Yep, it is. :-) Y1: 12200160415121876738 Y0: -19740274219868223167 and leave m out. Greg ... If so, how ...
                  Message 8 of 10 , Aug 12, 2005
                    Yep, it is. :-)

                    Y1: 12200160415121876738
                    Y0: -19740274219868223167

                    and leave m out.

                    Greg

                    --- In ggnfs@yahoogroups.com, Bouk de Water <bdewater@y...> wrote:
                    > Answering my own question perhaps :-)
                    >
                    > It's Y0 and Y1 isn't it, the non-monic parameters?
                    >
                    > Bouk.
                    >
                    > --- Bouk de Water <bdewater@y...> wrote:
                    >
                    > >
                    > > Hello everybody!
                    > >
                    > > I have a question.
                    > >
                    > > Is it possible to use a second irreducible polynomial with ggnfs?
                    If so, how
                    > > do
                    > > I enter it in the *.poly file?
                    > >
                    > > Say I wanted to use these two polynomials for a Lucas number:
                    > >
                    > > f(X) = 6*X^6 + 27*X^5 + 45*X^4 + 15*X^3 + 30*X^2 - 3*X + 4
                    > > g(X) = 12200160415121876738*X - 19740274219868223167
                    > >
                    > > with:
                    > >
                    > >
                    >
                    m=128812498419496899286734041447037910680019851818697424172503181354467240508369208475713248057682151827591523707537;
                    > >
                    > > How do I proceed? (if possible)
                    > >
                    > > Bouk de Water.
                    > >
                    > >
                    > > __________________________________________________
                    > > Do You Yahoo!?
                    > > Tired of spam? Yahoo! Mail has the best spam protection around
                    > > http://mail.yahoo.com
                    > >
                    > >
                    > >
                    > >
                    > > Yahoo! Groups Links
                    > >
                    > >
                    > >
                    > >
                    > >
                    > >
                    > >
                    >
                    >
                    > __________________________________________________
                    > Do You Yahoo!?
                    > Tired of spam? Yahoo! Mail has the best spam protection around
                    > http://mail.yahoo.com
                  • Bouk de Water
                    Leaving m out is indeed a good idea :-) - SNFS_DIFFICULTY is about 679.437899273151143712210122884678133053. Ouch. Better now! Thanks! Bouk. ...
                    Message 9 of 10 , Aug 12, 2005
                      Leaving m out is indeed a good idea :-)

                      -> SNFS_DIFFICULTY is about 679.437899273151143712210122884678133053.

                      Ouch.

                      Better now!

                      Thanks!

                      Bouk.

                      --- gchil0 <jgchilders@...> wrote:

                      > Yep, it is. :-)
                      >
                      > Y1: 12200160415121876738
                      > Y0: -19740274219868223167
                      >
                      > and leave m out.
                      >
                      > Greg
                      >
                      > --- In ggnfs@yahoogroups.com, Bouk de Water <bdewater@y...> wrote:
                      > > Answering my own question perhaps :-)
                      > >
                      > > It's Y0 and Y1 isn't it, the non-monic parameters?
                      > >
                      > > Bouk.
                      > >
                      > > --- Bouk de Water <bdewater@y...> wrote:
                      > >
                      > > >
                      > > > Hello everybody!
                      > > >
                      > > > I have a question.
                      > > >
                      > > > Is it possible to use a second irreducible polynomial with ggnfs?
                      > If so, how
                      > > > do
                      > > > I enter it in the *.poly file?
                      > > >
                      > > > Say I wanted to use these two polynomials for a Lucas number:
                      > > >
                      > > > f(X) = 6*X^6 + 27*X^5 + 45*X^4 + 15*X^3 + 30*X^2 - 3*X + 4
                      > > > g(X) = 12200160415121876738*X - 19740274219868223167
                      > > >
                      > > > with:
                      > > >
                      > > >
                      > >
                      >
                      m=128812498419496899286734041447037910680019851818697424172503181354467240508369208475713248057682151827591523707537;
                      > > >
                      > > > How do I proceed? (if possible)
                      > > >
                      > > > Bouk de Water.
                      > > >
                      > > >
                      > > > __________________________________________________
                      > > > Do You Yahoo!?
                      > > > Tired of spam? Yahoo! Mail has the best spam protection around
                      > > > http://mail.yahoo.com
                      > > >
                      > > >
                      > > >
                      > > >
                      > > > Yahoo! Groups Links
                      > > >
                      > > >
                      > > >
                      > > >
                      > > >
                      > > >
                      > > >
                      > >
                      > >
                      > > __________________________________________________
                      > > Do You Yahoo!?
                      > > Tired of spam? Yahoo! Mail has the best spam protection around
                      > > http://mail.yahoo.com
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      > Yahoo! Groups Links
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >


                      __________________________________________________
                      Do You Yahoo!?
                      Tired of spam? Yahoo! Mail has the best spam protection around
                      http://mail.yahoo.com
                    • trilliwig
                      ... Heh, just FYI, the latest CVS snapshots of GGNFS fix this little problem. You can specify or not specify m at your preference for non-monic linear
                      Message 10 of 10 , Aug 15, 2005
                        --- In ggnfs@yahoogroups.com, Bouk de Water <bdewater@y...> wrote:
                        > Leaving m out is indeed a good idea :-)
                        >
                        > -> SNFS_DIFFICULTY is about 679.437899273151143712210122884678133053.
                        >
                        > Ouch.
                        >
                        > Better now!

                        Heh, just FYI, the latest CVS snapshots of GGNFS fix this little
                        problem. You can specify or not specify m at your preference for
                        non-monic linear polynomials and the SNFS difficulty calculation
                        should work properly regardless. :)

                        --
                        Sam
                      Your message has been successfully submitted and would be delivered to recipients shortly.