init observations

[Paul A Houle]

Jeff Spaleta wrote:

>Is there some reasonably valid understanding of how big an impact this
>makes in the average and worst case? Sort of a measure of the spread
>of disk-drive performance out in the wild.
>Say for example...a typical christmas special home desktop from
>Dell... does the disk-seek performance on a system like that make an
>order of magnitude faster bootup or application startup times a
>pipedream?
>  
>
    My back-of-the-envelope number is 3x better performance in Ultra 320 
drives vs PATA for seek-heavy workloads (database apps,  make -j4,  cp a 
50 GB directory.)  A big part of the advantage comes from tagged command 
queueing -- recent SATA drives support Native Command Queueing (NCQ) but 
benchmarks I've seen of NCQ drives aren't impressive.