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

Re: [PrimeNumbers] ECM

Expand Messages
  • Andy Steward
    ... I agree. These days, I use B1=80*(Trial^(5/4)) and a random discriminant. Anyone wanting to use that UB code should replace those two lines with:
    Message 1 of 6 , Apr 24, 2002
    • 0 Attachment
      Walter Nissen wrote:

      > Andy Steward writes, in ECMSLIDE.UB:
      >
      > 12010 if EC% <= 25 then L1 = 2000 else if EC% <= 115 then L1 = 11000
      > else if EC% <= 415 then L1 = 50000 else if EC% <= 1115 then L1 =
      > 250000 else if EC% <= 2915 then L1 = 1000000 else if EC% <= 7015
      > then L1 = 3000000 else if EC% <= 17615 then L1 = 11000000 else L1
      > = 43000000 endif
      > 12300 L2 = L1*100

      > But, automatic incrementation seems optimal. The increment should
      > itself increment for each curve.

      I agree. These days, I use B1=80*(Trial^(5/4)) and a random discriminant.
      Anyone wanting to use that UB code should replace those two lines with:
      L1=ROUND(80.0*((1.0*EC%)^1.25)):L2=L1*100:SI=32768*IRND:SI=SI+IRND

      Mostly(*), I use GMP-ECM. I have to create batches specifying B1 for
      each trial as it's not a smooth increment.

      Andy

      (*) GMP-ECM is about three times faster than UB-ECM but requires a hard
      disk and Windows(**). Four of my oldest PCs have diskette only.

      (**) OK, I could set up a minimal Linux boot diskette but I haven't
      got around to it yet.
    Your message has been successfully submitted and would be delivered to recipients shortly.