improve strlen

hi all

i have read some manual pent opt;
i have tried write my pesonal routine faster than gcc compiler,
simply routine only
about strlen the result is the same time
my code and code is similar code found in paul hsieh page
i have tried read at the same time word and dword but result don't
chage
my routine work at the same time as strlen

any one know other solution (maybe mmx ist4uction) to improve the
algorithm)

regards

claudio


void RoutineC ( void )
{
char *source = "01234567890ABCDEF0123456789ABCDEF" \
"01234567890ABCDEF0123456789ABCDEF" \
"01234567890ABCDEF0123456789ABCDEF" \
"01234567890ABCDEF0123456789ABCDE\0" ; // 128 caratteri
unsigned int len = 0;

len = strlen ( &source[0] ) ;
}
void RoutineASM ( void )
{
char *source = "01234567890ABCDEF0123456789ABCDEF" \
"01234567890ABCDEF0123456789ABCDEF" \
"01234567890ABCDEF0123456789ABCDEF" \
"01234567890ABCDEF0123456789ABCDE\0" ; // 128 caratteri
unsigned int len = 0;

__asm__ ( " .p2align 4,,15\n\t"
" movl %%ebx ,%%edx\n\t"
"l1: movb (%%ebx) ,%%ah\n\t"
" incl %%ebx\n\t"
" test %%ah ,%%ah\n\t"
" jne l1\n\t"
" subl %%ebx,%%edx\n\t"
: "=b" (len)
: "d" (&source[0])
: "ebx", "edx"
) ;
len = len ;
}

.