Question about gmp-chudnovsky.c

richard rgoodmas at somtel.com
Tue Sep 12 07:35:54 CEST 2006


Paul Zimmermann wrote:

>>Date: Mon, 11 Sep 2006 11:35:01 -0400
>>From: richard <rgoodmas at somtel.com>
>>
>>When I run the gmp-chudnovsky program, I get a segfault at around 
>>8,000,274 digits.
>>Should this be expected on an old Athlon, 256K cache and 1.5G ram,  or 
>>is it more likely I have misconfigured something?
>>I'm running Debian gnu/linux, kernel 2.6.8-2-k7
>>Thanks in advance, Richard.
>>    
>>
>
>Did you consider item Q3 from the GMP FAQ at www.swox.com/gmp/#FAQ?
>Otherwise, you might consider increasing the stacksize limit (ulimit
>or unlimit depending on your shell).
>
>Paul
>
>  
>
Thanks for the prompt and helpful replies Paul, Blake, and Laurent.
As root, `ulimit -S -s 131072'  allowed `gmp-chudnovsky 100000000' to 
complete.

I will try recompiling with the enable-alloca flags, as default user 
stack is 8192K.
I'm using gmp-4.2.1 but haven't done any patching, as no test returned 
error.
I guess the x86-fat.diff patch can't hurt anything.  I need the 
compiling practice as
I'm not yet comfortable with the `new' automake, and its many options.
 
Thanks again,
Richard


More information about the gmp-discuss mailing list