Need help running ecm.exe from commandline
 

I am a TOTAL noob at this. I was hoping Wblipp(sp?) would give me some basic instructions, but he just gave me some composites.

I have NO IDEA how to run ecm without using a server. I simply need some basic instructions on running it from a command line, although hopefully I can put those commands in a batch file so my cpu doesn't run out of work when I'm not around.

Any help is tremendously appreciated.

[QUOTE=jasong]I am a TOTAL noob at this. I was hoping Wblipp(sp?) would give me some basic instructions, but he just gave me some composites.

I have NO IDEA how to run ecm without using a server. I simply need some basic instructions on running it from a command line, although hopefully I can put those commands in a batch file so my cpu doesn't run out of work when I'm not around.

Any help is tremendously appreciated.[/QUOTE]
First, run "ecm.exe -h" and read the help message the program gives you. it is probably worth capturing that output into a file so you can study it at your leisure.

After reading the help message (and, in my opinion, [i]only[/i] after reading it) create a file which contains the composite numbers you wish to factor --- one number per line.

Then choose a value for B1 appropriate for the size of factors you wish to find --- say, 3000000 for 40-digit factors --- and the number of curves you wish to run on each composite (let's choose 50 for an example). Then type the command

ecm -c 50 3000000 < inputfile..txt > outputfile.txt

and sit back and wait. Any factors will accumulate in the output file.


Modify the above command according to taste and to experience and, especially, when you've read and understood the help file enough to have an idea as to what would be useful.


Paul

Thanks Xilman, it's working now.:goodposting:

ok my turn.

I just compiled ecm.exe and tried to run ecm.exe -h but the window opens and closes in a blink what am I doing wrong?

Open a command line window first by typing "cmd" in the Start->Run dialogue box.

In the command line window you can run ecm.exe.

Alex

[QUOTE=akruppa;86366]Open a command line window first by typing "cmd" in the Start->Run dialogue box.

In the command line window you can run ecm.exe.

Alex[/QUOTE]

Ok thanks, Well now that I have read it and post after post on how to get started I'm even more confused but I'll keep trying. I just want to learn how to do & understand the cunningham tables. If someone can pick out a easy one so I can see how the whole process is done would be very welcomed. Hey we all started sometime LOL!!!

[QUOTE=grobie;86381]Ok thanks, Well now that I have read it and post after post on how to get started I'm even more confused but I'll keep trying. I just want to learn how to do & understand the cunningham tables. If someone can pick out a easy one so I can see how the whole process is done would be very welcomed. Hey we all started sometime LOL!!![/QUOTE]

What would you like to understand about the Cunningham Tables?
How they got started? Their history? Methods used to create the tables?

The best place to start is, of course, by reading the book:

[url]http://homes.cerias.purdue.edu/~ssw/cun/index.html[/url]

It not only discusses the theory and methods, but also contains
some very pleasant and interesting reading about the early days of
computing.

Reading an elementary number theory book might also be beneficial.
I can recommend some if you like. It is a requirement before you can
even start learning about methods. One has no hope of understanding
even the beginnings of prime testing methods without knowing what a
primitive root is, for example. One can't start learning about factoring
methods without knowing what a quadratic residue is. etc. etc.

Is this the book you are talking about?

[url]http://www.ams.org/online_bks/conm22/[/url]

Yep, that's the Cunningham book.

Alex

Well I guess I'll just stick with what I know, this is a little over my understanding. Or I'll just use the ecmclient :surrender

Just flip through some easy introductory texts on number theory and pick one you like. Start with easy stuff. As Bob points out, the sub-exponential factoring algorithms are hard to understand without knowing at least some basics in number theory.

Start, for example (off the top of my head), with residue classes, the congruence (mod n) relation, co-primality and Euler's phi function, multiplicative inverses, perhaps quadratic residues and quadratic reciprocity.

Alex