DDJ article / UNIX vs BS/2

[William E. Davidsen Jr]

In article <209 at imspw6.UUCP> bob at imspw6.UUCP (Bob Burch) writes:

| There is a simple and deadly counter-argument to any and all of this and,
| again, you don't need to be Albert Einstein to figure it out:
|  
| At any point in time, you will buy yourself some gain in performance
| going with a proprietary OS versus UNIX for a given piece of hardware
| which, presumably, wasn't specifically built to run UNIX (something like
| one of the Gould "firebreathers" which WAS specifically blueprinted for
| UNIX is a different story).  I don't know exactly what the performance
| gain is for a typical VAX or HP mini but, for the sake of argument, let's
| assume it is 50 percent, which I suspect is being generous.  So you and I
| each buy one such computer at the same time, mine with UNIX, yours with
| the proprietary OS, and you've got me by 50%.

  Unfortunately, at least on a VAX, your argument does not seem to match
the observed results. I talked to four system managers who handle both
VMS and UNIX VAXen, and even those who really dislike UNIX as an
interface agree that UNIX will run 20-30% more users on a VAX, at least
when doing typical things like edit, compile, read mail, light
computation, etc.

  The reason is that VMS has a lot of overhead in starting a process,
and a lot in file i/o, due to the many types of file. You have some
features added in VMS (which may or may not be needed), but you pay for
them.

  I will agree with you that it is possible to write an o/s for any
given computer which will maximize performance, but it is not a given
that performance is the goal of a proprietary o/s. In fact, given the
low overhead of process startup and file i/o in UNIX, there is usually a
limited place for improvement there. Using the fast file system (BSD and
V.4) the overhead of directory access in UNIX is low.

  The vendors want to have features which (a) the user can understand,
and (b) which lock the user into the o/s. If UNIX has an overhead of 30%
(pick any small number), the most a vendor could gain would be that
overhead, even with a perfect o/s. In practice most applications don't
have even that much overhead, but that's VERY application dependent.

  This doesn't mean that UNIX is perfect for every application and
machine, but it suggests that a vendor will add features rather than
performance to make his product appeal to users (or actually management,
which isn't the same thing at all).

  To repeat a few of your lines:  
| At any point in time, you will buy yourself some gain in performance
| going with a proprietary OS versus UNIX for a given piece of hardware
| ...
  The assumption that proprietary=faster is not universally true,
although I agree that for any machine there is room to custom tailor the
software to the hardware. Perhaps the decision of Cray and DEC to go
with UNIX for new machines indicates that portability is important also,
and that custom tailored systems are often so expensive to port to new
architecture that it is impractical. It could be argued that this may
actually slow the development of new hardware, but I have nothing to
back that other than gut feeling.
-- 
	bill davidsen		(wedu at ge-crd.arpa)
  {uunet | philabs}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me