I had a look at the numbers you got on the benchmark you posted, and I've
tried averaging out the values to see if -Rstripe made any difference.
FWIW, it seems there's a small (but existant, IMHO) improvement on reads
and not much change on writes. Reads seemed more improved running 2+
threads, perhaps because less than that is quite slow. I calculated
averages over the various chunk-sizes (since I'm ignoring their effect
anyway), and used the numruns 3 numbers which look more consistent.
If anyone wants an xspread or dump of the numbers, let me know.
Cheers,
--
_/\ Christian Reis is sometimes [EMAIL PROTECTED]
\/~ suicide architect | free software advocate | mountain biker
