HP 3000 980/100 beats the IBM 3090 180J at one-third the cost !!!

A recent batch order processing benchmark of a large company
demonstrates that the HP 3000 980/100 can beat an IBM 3090 180J both in
performance and five year cost of ownership. The following is a
description of the benchmark and a slide that can be used in customer
presentations. For more information a Performance Brief "HP 3000
Mainframe Batch Performance Benchmark" is available from LDC PN# 5091-
2553E.

Benchmark Description

Systems

o IBM 3090 180J water-cooled mainframe configured with MVS/ESA SP
  version 3.1.3, 128 Megabytes of central storage, 128 Megabytes of
  expanded storage, and 3380/3990 DASD technology.

o HP 3000 Series 980/100 configured with an enhanced version of MPE/XL
  release 3.0 (available 1H92), 640 Megabytes of main memory, and HP
  6000 670 fiber-link disk drives.

Benchmark

This benchmark executes an order processing job stream that includes
the sorting of a million order records.  The benchmark is based on an
actual daily order processing job from a large company.  It was ported
from MVS to the HP 3000.

Results - Benchmark and System Cost Analysis

The HP 3000 had better throughput running the benchmark in 14.5 minutes
versus 15.6 minutes on the IBM 3090.  The HP 3000 also executed 60%
fewer disk I/Os than the mainframe.  The ability to minimize disk I/O,
and reduce the amount of time waiting on I/O, is a crucial factor in
providing responsive systems for critical application processing.

The HP 3000 980/100 had a five year cost of ownership of $1,068,000
versus the IBM 3090 180J  five year cost of $3,501,000!  In fact, the
maintenance cost alone on the IBM system could purchase and maintain
the HP system for five years.