🚀 go-pugleaf

I have recently concluded some system tests on disk performance.  Included in
these tests have been some newer low end workstations.  One of the results is
that "SCSI" is slow.  In fact, I think I have some new nominations for the
"World's Fastest Toy Computer" Award.  Two questions:

1) Why are SCSI disk subsystems *so* slow.  On sequential reads, ~300 KB/sec.
is typical.  The filesystems and all hardware interfaces should be able to
easily sustain 3X as much on sequential reads of large, unfragmented files.  
(Which SCSI disk subsystems you ask?  I would rather turn it around, and say,
are there any exceptions to the above sweeping generalization?  :-)  I use
SMD disks as a baseline...)

2) Is the reason lack of gather/scatter on (inexpensive) SCSI controllers?
What companies use gather/scatter SCSI controllers?  What companies
manufacture fast synchronous SCSI controllers with gather/scatter?  Will the new
Synchronous SCSI controllers have gather/scatter, in general?  Why does 
gather/scatter matter, anyway - it looks like an extra block copy should be 
a trivial cost on most new systems (wrt CPU and memory bus cycles).

3) Are there implementation limitations of the new RISC-based systems which 
make I/O cost more than on older systems?  (e.g. VAX or 68K).  Is there
something about SCSI in particular that is a problem?

4) Are there any architectural limitations of the new systems which might be
causing a problem?   (I am thinking of architectural limitations that might
affect cacheing and memory mapping that would indirectly require more trips
through the device drivers, for example).

5) (Not really a comp.arch question, but related to the above - aside:
Does anyone make synchronous SCSI disks which really perform?)

  Hugh LaMaster, m/s 233-9,  UUCP ames!lamaster
  NASA Ames Research Center  ARPA lamaster@ames.arc.nasa.gov
  Moffett Field, CA 94035     
  Phone:  (415)694-6117