> > Should I be bencmarking against the memcpy() in lib/string.c,
> > or the one in include/asm/string.h ?
>
> The latter; the one in lib is used if the architecture doesn't have an
> optimised memcpy() routine. For some reason, people seem to consider
> this an important optimisation :-)
Ok, corrected benchmarks:
lib/string.c (bytecopy)
Cached 2m 18s
Non cached 2m 43s
asm/string.c (rep mov)
Cached 24s
Non cached 31s
MMX
Cached 16s
Non cached 25s
Seems that the MMX copy wins in both cached and non-cached situations.
regards,
Dave.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/