20071221, 11:28  #1 
Just call me Henry
"David"
Sep 2007
Cambridge (GMT/BST)
5·31·37 Posts 
Curves needed
While using the information on http://www.loria.fr/~zimmerma/records/ecm/params.html with gmpecm 6.1.3 i noticed that v6.1.3 wanted less curves run at most levels
why then are we still using the v6.0 amount of curves Digits v6.0 v6.1.3 Diff 20 77 74 3 25 206 214 8 30 401 430 29 35 948 904 44 40 2,440 2,350 90 45 4,590 4,480 110 50 7,771 7,553 219 55 17,899 17,769 130 60 43,670 42,017 1,653 65 69,351 69,408 57 also how are these numbers calculated 
20071221, 14:08  #2 
Bamboozled!
"πΊππ·π·π"
May 2003
Down not across
3×11×307 Posts 
Because small changes in the number of curves makes essentially no difference to the amount of cpu time required to find factors.
Left as an exercise in the use of Google. Paul 
20071221, 15:47  #3 
Just call me Henry
"David"
Sep 2007
Cambridge (GMT/BST)
5·31·37 Posts 
surely 1600 curves at 60 digit level is worth missing

20071221, 16:13  #4 
Bamboozled!
"πΊππ·π·π"
May 2003
Down not across
3·11·307 Posts 
Read the numbers again.
Approximately 43K curves are recommended at this level. 1600/43K is under 4%. That in itself is barely worth worrying about. When you take into account (whiich you clearly haven't) that the extra work gives you a larger chance of finding a factor if one is to be found, then it really isn't worth making a fuss over it. Paul Last fiddled with by xilman on 20071221 at 16:14 Reason: Fix tyop. 
Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Number of curves needed?  jibanes  GMPECM  3  20180204 19:55 
ECM  why 3 curves?  James Heinrich  PrimeNet  3  20171114 13:59 
JKLECM: ECM using Hessian curves  CRGreathouse  Software  1  20170906 15:39 
B1 and # curves for ECM  Walter Nissen  Factoring  36  20140216 00:20 
Need help with elliptic curves...  WraithX  Math  12  20100929 09:34 