md5sum makes unaligned word accesses...

[ldoolitt at recycle.lbl.gov]

Guys -

On Sun, May 14, 2006 at 07:06:01PM -0500, Michael S. Zick wrote:
> On Sun May 14 2006 19:00, Rob Landley wrote:
> > On Friday 12 May 2006 6:21 pm, Mike Frysinger wrote:
> > > so, we're already penalizing systems, we just dont notice ;)
> > I'm talking about binary size, not performance.  And the big bottleneck is the 
> > cache line fetches, not the internal processing; try actually benching stuff 
> > and the alignment is likely to be noise at best.
> >
> Say what? Read:
> linux/arch/parisc/kernel/unaligned.c which is called from a trap handler.
> Are you saying that process is faster than a data cache line fetch?

No, Rob is talking about machines like the '386, where
the hardware fixes up the alignment.  Fixing the
alignment is either mandatory to get the right answer,
or a big performance win on other systems, like ARM and MIPS.
The performance improvement on a '386 is relatively small.

         - Larry