[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [leafnode-list] getaline() (once more)
krasel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx (Cornelius Krasel) writes:
> Matthias Andree wrote:
> > I did not see what was wrong with my getline/getaline implementation,
>
> IIRC you suspected that your getaline implementation was very slow.
I did some benchmarking on my machine (a Celeron with 400 MHz and 64 MB
RAM). Both versions of getaline had to read a textfile of a million lines,
each 60 chars long, via stdin. The 2.0b7 version of getaline needed between
4.0 and 5.0 seconds for this; the getaline version from 1.9.17ma3 took
between 10.9 and 12.0 seconds.
Here are some more details:
2.0b7:
real 0m4.002s 0m4.006s 0m4.997s 0m4.102s 0m4.224s
user 0m2.020s 0m1.860s 0m2.070s 0m1.980s 0m1.900s
sys 0m0.350s 0m0.420s 0m0.360s 0m0.400s 0m0.500s
1.9.17ma3:
real 0m10.909s 0m11.227s 0m11.991s 0m10.942s 0m11.013s
user 0m9.700s 0m9.580s 0m9.580s 0m9.630s 0m9.640s
sys 0m0.300s 0m0.450s 0m0.450s 0m0.370s 0m0.350s
The program was:
int main( void ) {
char *l;
int i = 0;
while ((l = getaline(stdin))!= NULL) {
i++;
}
printf("%d\n",i);
return 0;
}
I suspect one could live with a loss of about 60% of speed since the
"ma" version still is able to read a line in approx. 11 microseconds.
The bottleneck is probably not the speed of getaline() but the speed of
the network connection, and this will even be true for a 486.
This leaves the question of licensing. Matthias' getaline() is licensed
under the GNU Public license. This would require to put leafnode-2.0
also under the GPL. While I am not a close friend of the GNU foundation,
I don't have problems with the GPL but I wonder what other people on
this list think.
--Cornelius.
--
/* Cornelius Krasel, U Wuerzburg, Dept. of Pharmacology, Versbacher Str. 9 */
/* D-97078 Wuerzburg, Germany email: krasel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx */
/* "Science is the game we play with God to find out what His rules are." */
--
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list