disks=$(awk ‘/sd[a-z]+$/{print $4}’ /proc/partitions );
echo creating output for $disks…
for i in $disks; do smartctl –all /dev/$i > $i.out & done
wait;
echo checking for ATA Err string
for i in $disks; do
grep “ATA Err” $i.out
done
echo done

If you’re not sure about the performance of your disks, it’s always worth doing a quick check or two with dd(1):

dd if=/dev/sda of=/dev/null bs=1M count=16384

This will read a gig off the beginning of the disk and tell you the read rate at the end. Thats likely to be the fastest you could ever, under the most optimum conditions, read from this disk. You can also read from a RAID device (e.g., /dev/md1) to see how much RAID is costing you compared to the throughput available from individual disks. Writes can be done with “if=/dev/zero of=/dev/whatever conv=fdatasync”, though keep in mind that that wipes out data! If you already have a filesystem, you probably just want to write to a file in that filesystem instead, though this gives you the speed after filesystem (and perhaps RAID) overhead.