- One additional "enhancement" of 5.0c vs 4d-r7a I left out, was that

5.0c now "maintains" the expressions. Here is some "sample" output

c:\ecm\ecm-5.0c>echo 27.691#+1 | ecm -c 3 -ve 200 -deep 15000

GMP-ECM 5.0c [powered by GMP 4.1] [ECM]

Input number is 27.691#+1 (291 digits)

Using B1=15000, B2=2114895, polynomial x^1, sigma=2214210635

Step 1 took 2313ms

Step 2 took 1522ms

Line=1 Curves=2/3 B1=15000 Factors=0

Input number is 27.691#+1 (291 digits)

Using B1=15000, B2=2114895, polynomial x^1, sigma=579468029

Step 1 took 2414ms

Step 2 took 1532ms

********** Factor found in step 2: 3302513916338587

Found probable prime factor of 16 digits: 3302513916338587

Composite cofactor

2270478349675689184681254371761976795334832608365966034800675383016042

5494115265110948714279402342416540729753940671497012750001602757770976

4024230872996914811635410193141366258313816663972516000113519668852437

273257579884363472903020309122119120061184840304026980285988591553

has 276 digits

Line=1 Curves=3/3 B1=15000 Factors=1

Input number is (27.691#+1 )/3302513916338587 (276 digits)

Using B1=15000, B2=2114895, polynomial x^1, sigma=974840061

Step 1 took 2163ms

Step 2 took 1422ms

Line=1 Curves=4/3 B1=15000 Factors=1

The line "Input number is (27.691#+1)/3302513916338587 (276 digits)

shows this "keeping" of expressions. However, this small example

has also shown me that I missed one (and need to fix it). The

original output of the cofactor is fully expanded, when the

"expression" could have been output.

Jim. - --- In primenumbers@yahoogroups.com, "Jim Fougeron." <jfoug@k...>

wrote:> One additional "enhancement" of 5.0c vs 4d-r7a I left out, was that

Thanks for adding these features, especially the factorial and

> 5.0c now "maintains" the expressions. Here is some "sample" output

>

>

> The line "Input number is (27.691#+1)/3302513916338587 (276 digits)

> shows this "keeping" of expressions. However, this small example

> has also shown me that I missed one (and need to fix it). The

> original output of the cofactor is fully expanded, when the

> "expression" could have been output.

>

> Jim.

multifactorial code! The idea of "keeping" expressions will also be

very handy. I've run recently large files for n!+-1 for 600<n<1000

Searching through the output for all the factors was a right pain!!

I've probably made many typos in my lists (at

www.uow.edu.au/~ajw01/curves/ the plus and minus files.)

This will at least help check them all. Does the code sort the factors

in this output?

One other possibility worth considering is adding a definitive

primality test for small factors and cofactors. It should be off by

default with an option to select the maximum digit size for which it

is called. Of course if the AKS conjecture 4 ever gets proved, this

will be feasible for quite decent sizes!

Andrew