Performance Units, for a particular CPU, translated to relative performance
between two CPUs needs to be used with care. There are the obvious other factors
that will affect throughput (isn't that what it's all about?). Disk I/O and
memory.

We recently tested a [third-party data-base tool] process on a test machine, did
the arithmetic for the production machine and expected a pro-rata throughput
increase. It did not happen, and the reason, we believe, is that the disk drives
on the production machine mitigated the expected overall performance gain. In
our particular case, the fact that the production machine is a 997/1200 while
the test machine is a 979/3000, attributes, specifically number of processors,
that are factored into performance units were irrelevant to our  [third-party
data-base tool] process.

Where performance units may be used with greater confidence, I believe, is
overall throughput.

I'd be eager to be corrected on this.

===================
Leonard S. Berkowitz
Perot Health Care Systems
(Harvard Pilgrim Health Care account)
voice: 617-509-1212
fax:   617-509-3737
pager: 781-226-2431







Richard Hoffmann <[log in to unmask]> on 12/19/2000 10:36:56 AM

Please respond to [log in to unmask]








 To:      [log in to unmask]

 cc:      (bcc: Leonard Berkowitz/CORP/HPHC)



 Subject: [HP3000-L] Performance Units for 987/200








Can someone give me the Performance Units for a 987/200.   I have the
number relative to a 918 which is 6.0, but I cant find the other number.

Thanks
Richard



--
Richard Hoffmann                                 http://www.blinncol.edu

Systems Analyst/Programmer                 979-830-4213  (Work)
Blinn College                                         979-830-4049
(Fax)
Brenham, Texas