MSCP SCSI controller speed - test-drb@ccmp.vtda.org

9 Nov 2005

...

Subject: Re: MSCP SCSI controller speed
   From: Paul Koning <pkoning at equallogic.com>
   Date: Wed, 09 Nov 2005 11:22:02 -0500
     To: cctalk at classiccmp.org

 Allison> You need a real time or single user os to get better
 Allison> results.  Plain old dos might do far better or maybe a
 Allison> custom kernal tuned for this use.

All you need is an application, or OS, that can queue up multiple
transfers.  I thought the buffered I/O in Unix would take care of
that.

Certainly if you do single I/Os and wait for the answer each time,
things will be slow, but the OS isn't really at fault then.

       paul 
True.  However a RT OS has better known and predictable latency.
There is also the issue of OS overhead for process control.  

Another way this could be improved even in the Netbsd case is to 
implement LRU caching so that when the IO request appears the need 
to hit the device for a transfer is defered.  If the buffer is large 
the likelyhood of a cache hit is high and the idel time process can 
be writeback and read ahead.

Allison