Re: gmp issues (long)

[Rob Browning]

Stephen Compall <address@hidden> writes:

> Rob Browning wrote:
>      replace the _mp_size field with a short, since I doubt that we'd
>      ever be likely to overflow long data[MAX_SHORT] digits. (What was
>      the size limit, if any, on our old bignums?).
>
> My first use of GMP in C (after the requisite playing around while 
> learning) was to try to build up a Godel number as an example in a 
> project involving semigroups I'm working on (OT). After creating one of 
> these numbers, I believe the memory usage of the mpz_t & friends was ~5 
> MB. While this was an abnormal case, it's still far off from 256k. 
> er....

Wow.  That's a pretty giant number :>

OK.  If we want to accomodate numbers this big without a 2-tier hybrid
approach, or something, people should adjust my storage computations
to account for a 4-byte size rather than a 2-byte-size.

Also, just for comparison, and presuming I read the code correctly,
our previous implementation used a 16-bit (plus sign bit) size field,
so moving to a 4-byte size field would be a substantial increase over
what we supported before.

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG starting 2002-11-03 = 14DD 432F AE39 534D B592  F9A0 25C8 D377 8C7E 73A4