Re: [tlug] SATA software RAID or SAS hardware RAID?

["Stephen J. Turnbull" <stephen@example.com>]

Edward Middleton writes:

 > You might get more of a speed increase by reformulating your problems so
 > that it is more distributable.

You're still missing the point.  The computation is perfectly
distributable; each subtask is entirely independent of all the others.
The bad performance is entirely that 4 CPUs sharing two memory banks
spent more time *waiting* for each *single* task to report than it
took 1 CPU to do 1000 such tasks.