memcpy performance
Joel Sherrill
joel@OARcorp.com
Tue Dec 9 08:57:00 GMT 1997
On Tue, 9 Dec 1997, Jeffrey A Law wrote:
>
> In message <199712091644.IAA20455@cygint.cygnus.com>you write:
> > An RTEMS user has noticed that the performance of the RTEMS message
> > queue routines is bound by our use of newlib's memcpy routine. The
> > current implementation uses a byte-by-byte copy.
> >
> > What would it take to get a better generic implementation? Is there one
> > laying around which could be used instead? I remember the old GNU/960
> > version did some alignment and then copied 16-bytes at a time.
> Already done :-)
>
> Tue Oct 7 14:01:29 1997 Bill Moyer (billm@cygnus.com)
> Jeffrey A Law (law@cygnus.com)
>
> * libc/string/{memchr.c, memcmp.c, memcpy.c}: Reimplement to run faster.
> * libc/string/{memset.c, strcat.c, strchr.c}: Likewise.
> * libc/string/{strcmp.c, strcpy.c, strlen.c}: Likewise.
> * libc/string/{strncat.c, strncmp.c, strncpy.c}: Likewise.
Great!!! Please feel free to change the RTEMS configuration to prefer
speed over size. :)
--joel
More information about the Newlib
mailing list