[QUOTE=Andi47;172573]the windows binary for ggnfs found [URL="http://gilchrist.ca/jeff/factoring/"]here[/URL] won't run on my office PC (Pentium 4, Win XP pro), it crashes immediately with an unknown software exception (0xc0000094) at position 0x00418b8d.[/QUOTE]

*grrrr* - why doesn't ggnfs give more precise error messages? I just found out, that the inputfile was missing the "n: <number>" line, now, after adding this line, it is running.

[QUOTE=Andi47;172676]*grrrr* - why doesn't ggnfs give more precise error messages? I just found out, that the inputfile was missing the "n: <number>" line, now, after adding this line, it is running.[/QUOTE]The message you're getting is not a GGNFS error. It's a Microsoft error. I've said this before. Just disable the error reporting.

[QUOTE=smh;172679]The message you're getting is not a GGNFS error. It's a Microsoft error. I've said this before. Just disable the error reporting.[/QUOTE]

It is not a microsoft error. It is a division by zero exception.

(I'm trying to rebuild the program with debug symbols to pin it down.)

Here's a gdb backtrace.

[code]#0 0x00007f851d119bb4 in __gmp_exception (error_bit=2) at errno.c:44
#1 0x00007f851d119be2 in __gmp_divide_by_zero () at errno.c:60
#2 0x00007f851d14ec01 in __gmpz_tdiv_r (rem=0x7fff25410cf0,
num=0x7fff25410cf0, den=0x62bad0) at tdiv_r.c:41
#3 0x00007f851d1432a9 in __gmpz_mod (rem=0x7fff25410cf0,
dividend=0x7fff25410cf0, divisor=0x62bad0) at mod.c:45
#4 0x000000000040b961 in input_poly (N=0x62bad0, A=0x62bb30, adeg=0x6b9c90,
B=0x62bb38, bdeg=0x6b9c94, m=0x62bac0, fp=0x2199370) at input-poly.c:211
#5 0x00000000004056a3 in main (argc=<value optimized out>,
argv=0x7fff25411958) at gnfs-lasieve4e.c:607
[/code]

Oh that was with the "experimental" version of ggnfs, so the line numbers may be different if you're using a different version.

But basically input_poly is not sanity-checking.

I wouldn't worry about it. It's user error (sorry :razz:)

[QUOTE=jrk;172685]Oh that was with the "experimental" version of ggnfs, so the line numbers may be different if you're using a different version.

But basically input_poly is not sanity-checking.

I wouldn't worry about it. It's user error (sorry :razz:)[/QUOTE]

I know, that this is a user error - but when ggnfs would sanity-check the input and give a precise error message, it would be easier to detect the error. (not all input errors are THAT obvious)

You are right.

[quote=jasonp;170971]If A(x) = A5 x^5 + A4 x^4 + ... + A0 is the algebraic polynomial, then pretend x = a/b, where these are the coordinates in the relation, so that A(a,b) = b^5 *A(a/b) is the homogeneous form of the polynomial. Basically you multiply the coefficient A_i by a^i * b^(5-i).[/quote]
I can't get this to work. I even tried coming back to it when i was less tired to try again which often works.
Could i bother you for an example?

[tex]f(x)=x^5+2x^4-2x^3+3x^2-4x+5[/tex]
[tex]F(a,b)=b^5*f(\frac{a}{b}) = b^5*(\frac{a^5}{b^5} +2\frac{a^4}{b^4}-2\frac{a^3}{b^3}+3\frac{a^2}{b^2}-4\frac{a}{b}+5)[/tex]
Now multiply all the terms by [tex]b^5[/tex]
[tex]F(a,b)=b^5*f(\frac{a}{b})=a^5+2{a^4}b-2{a^3}{b^2}+3{a^2}{b^3}-4a{b^4}+5b^5[/tex]

Hopefully this helps.

[quote=joral;173083][tex]f(x)=x^5+2x^4-2x^3+3x^2-4x+5[/tex]
[tex]F(a,b)=b^5*f(\frac{a}{b}) = b^5*(\frac{a^5}{b^5} +2\frac{a^4}{b^4}-2\frac{a^3}{b^3}+3\frac{a^2}{b^2}-4\frac{a}{b}+5)[/tex]
Now multiply all the terms by [tex]b^5[/tex]
[tex]F(a,b)=b^5*f(\frac{a}{b})=a^5+2{a^4}b-2{a^3}{b^2}+3{a^2}{b^3}-4a{b^4}+5b^5[/tex]

Hopefully this helps.[/quote]
thanks
now i understand it i cant see why i couldn't understand it with the previous explanation:smile:
:pancakebunny:

[quote=henryzz;173112]thanks
now i understand it i cant see why i couldn't understand it with the previous explanation:smile:
:pancakebunny:[/quote]

Welcome to the club!!! (Of people who don't understand things when they're first explained to them.)

When I was first taught how algebraic division worked I didn't understand :smile: