Controversial opinion: I think mkdep.c is not speed critical enough to
warrant architecture-specific tests.
(I know all about the inner loop (in fact I added the MIN and MAX tests
to bum more speed off Linus's implementation), so if you don't agree
with me with my priorities, great, don't agree, but please, no "you
don't know enough to have an opinion" responses.)
Would someone like to try killing the LE_MACHINE stuff and change GETNEXT
to something like this:
if ( next > end )
break;
current = *next++;
Then run good benchmarks on a variety of machines, including a Pentium and
a Sparc. My guess is that the L1 cache will give some of the performance
of the existing scheme. I'm not enough of a cache guru to estimate how
much.
Michael Elizabeth Chastain
<mailto:mec@shout.net>
"love without fear"
-
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/