Performance

[Andreas Hasenack]

Em Tue, Nov 20, 2001 at 07:37:35PM +0100, Werner Koch escreveu:
> On Tue, 20 Nov 2001 13:41:06 -0200, Andreas Hasenack said:
> 
> > I ran this test again with openssl also recompiled for i686 and
> > got a small improvement:
> 
> Hmmm, I wonder why openssl should have any influence on gpg.
> 
>   Werner

Actually, the difference is very small, I ran it again (using assymetric
and symmetric):
openssl@i686, gnupg@i686:
$ time gpg -v -e -r andreas --cipher-algo RIJNDAEL lala
gpg: usando chave secundária 61D6A604 ao invés de chave primária F3DDC290
gpg: Esta chave pertence a nós
gpg: lendo de `lala'
gpg: escrevendo para `lala.gpg'
gpg: ELG-E/RIJNDAEL criptografado para: 61D6A604 Andreas Hasenack <andreas@netbank.com.br>

real    0m48.100s
user    0m30.640s
sys     0m2.110s


openssl@i386, gnupg@i686:
real    0m47.752s
user    0m31.230s
sys     0m2.100s

Probably something else started running on my machine when I ran the previously
tests, since gnupg is not linked against openssl :)

Comparing to plain i386:
openssl@i386, gnupg@i386:
real    0m53.866s
user    0m31.670s
sys     0m2.010s