> > 	bigger blocks, lower overhead,
> Uhm, that was my first thought too on this matter. However, when I started
> benchmarking with a kernel with Koji Imada's patches, the result was exactly
> opposite... The suggestion was that the drive's firmware might have been
> optimized for 512 byte/blocks. I must say that this was an old-drive, maybe
> modern drives to better.

In fact, most current drives no longer support changing the blocksize, because
Windows does not support anything other than 512 byte hard disk sectors (ta
da!).  Many slightly older drives had cache code optimized for (or in extreme
cases, written solely for) 512 byte sectors and thus performed poorly after 
reformatting.  The days when reformatting to remove sector gaps made sense are
probably long gone.  (Note that a lot of current IBM drives have exactly one
"sector" per track, and emulate 512 byte sectors only at the command interface
level; as long as the drive is caching the entire track...)