mersenneforum.org On the efficiency of SNFS for "dirty" numbers
 Register FAQ Search Today's Posts Mark Forums Read

 2018-09-29, 17:23 #1 didgogns   Mar 2014 South Korea 24 Posts On the efficiency of SNFS for "dirty" numbers Hi, This time I want to factor (2^300-1)*10^91+2^299-1 and similar numbers. The number is very close to 2^300*10^91 and a reasonable SNFS polynomial will be something like 10*x^5+2^299-10^91-1, where x=2^60*10^18. Now my concern is that c0=2^299-10^91-1 looks very big. The question is: Is there significant performance drop if SNFS polynomial is "dirty" like this case? If so, how the people who SNFSed 150^149+149^150 or other numbers solved this issue?
2018-09-29, 17:48   #2
didgogns

Mar 2014
South Korea

24 Posts

So the SNFS polynomials of XYYXF numbers are of the these form.
Quote:
 Originally Posted by XYYXF SNFS is faster but is restricted to numbers of special form, e.g. sextic a*b6+c*d6 or quintic a*b5+c*d5 for Leyland numbers.
How can I create a .poly file of these form? Is there any online guide?

2018-09-29, 18:22   #3
R. Gerbicz

"Robert Gerbicz"
Oct 2005
Hungary

26258 Posts

Quote:
 Originally Posted by didgogns The number is very close to 2^300*10^91 and a reasonable SNFS polynomial will be something like 10*x^5+2^299-10^91-1, where x=2^60*10^18. Now my concern is that c0=2^299-10^91-1 looks very big. The question is: Is there significant performance drop if SNFS polynomial is "dirty" like this case?
I'd call it uninteresting, digit related, totally boring numbers. But there is a better polynom, most likely this is still gnfs job. Hint: 2 | 10.

 Similar Threads Thread Thread Starter Forum Replies Last Post 2147483647 Factoring 2 2016-12-10 08:42 pinhodecarlos NFS@Home 2 2015-07-04 11:18 gszpetkowski Factoring 13 2014-08-05 11:57 nitai1999 Software 7 2004-08-26 18:12 antiroach Lone Mersenne Hunters 6 2003-07-16 23:35

All times are UTC. The time now is 03:11.

Thu Jan 28 03:11:34 UTC 2021 up 55 days, 23:22, 0 users, load averages: 3.70, 3.81, 3.42