Cray/Media:      Mardi Larson, 612/683-3538
Cray/Financial:  Brad Allen, 612/683-7395

CRAY RESEARCH ANNOUNCES REVOLUTIONARY NEW
SOFTWARE TOOL FOR PLASTICS, AUTOMOTIVE AND
ASSOCIATED INDUSTRIES

New Product Engineered In Cooperation With General
Motors Research & Development Center

DETROIT, Nov. 15, 1995 -- At the Society of Manufacturing
Engineers (SME) AutoFact 94 conference held here this week,
supercomputer leader Cray Research, Inc. announced a
significant new software product that it said could
revolutionize the way leading-edge plastics manufacturing is
done in the plastics, automotive and plastic-associated
industries, the company said.   The new software, Cray Molding
Logic (CMLogic) 2.0, was engineered in cooperation with
General Motors Research & Development, Cray Research said. 

Company officials said that CMLogic software predicts optimal
manufacturing solutions for plastic injection molding
processes.  It is expected to reduce the costs associated with
plastic injection molding manufacturing, help speed production
of plastic parts and components, and result in higher quality
products.  These new capabilities, the company said, will be
attractive to many industries that manufacture plastics goods,
especially the automotive industry, where today more than 50
percent of the end products is made of plastic and high
performing polymers. 
 
The new product combines supercomputing with classic
design-of-experiment methodology and industry-standard
plastic injection molding simulation programs for an
automated, predictive solution offered only on Cray Research
supercomputers.  In its prototype stage earlier this year,
CMLogic was used by several industrial companies to predict
the molding process setups for a variety of plastic parts
including a new cellular phone housing unit, an automotive
engine manifold, and an automotive instrument panel bezel,
according to Mike Obermier, a Cray Research polymer
processing specialist and head of the CMLogic development
team.  

Using CMLogic on Cray supercomputers, manufacturers can save
months in time-to-market because they can predict a mold's
productivity long before a molding tryout (physical test of
mold) begins, Obermier said.  "Molders can start mold tryouts
with complete confidence, and they can accelerate project
scheduled because they will need fewer mold tryouts. 
Ultimately they can produce improved parts with tighter
tolerances and higher quality at lower costs -- essential to a
company's global competitiveness today," Obermier said. 

According to John Ulicny, research engineer at General Motors,
a large automotive body panel mold can cost up to $500,000
and may require up to ten mold trials.  "Considering that a
single molding trial may cost as much as $40,000 and take up
to six weeks effort, using CMLogic to eliminate even one of
those trials can create substantial savings," he said. 

John Carlson, Cray Research chairman and chief executive
officer, said CMLogic represents a growth initiative for Cray
Research.  "We are leveraging our expertise in high-
performance simulation into new markets like the plastics
industry.  Our success in other commercial industries has been
primarily based on the value that simulation brings (over
physical prototyping methods) to time-to-market savings and
improved product quality.  This will be an important factor in
our success in taking our simulation tools to new markets in
the future," he said.

Obermier said plastic injection molding simulation is a
relatively new application for supercomputers that has been
embraced first by the automotive industry and plastics
material suppliers over the past few years.  The leading
injection molding simulation packages Moldflow and C-Mold
have been running on Cray Research supercomputers since
1991. 

"Today, with the announcement of CMLogic, we are taking this
plastic injection simulation technology one giant leap
forward," said Obermier.  "The simulation programs reliably
predict a mold's performance for a single molding machine
setup.  CMLogic software simulates a designed experiment and
runs multiple molding setups, integrates those runs, and
'automatically' predicts the range of acceptable machine
setups for a particular plastic product or part.  With CMLogic's
predictive capabilities, manufacturers can optimize molding
equipment machine cycle times, or quickly resolve current
production problems."  

According to Obermier, CMLogic is designed to work with
either Moldflow or C-Mold as the molding simulation code,
making it easy for customers and prospects currently using
these tools to move into CMLogic's advanced capabilities. 

CMLogic allows users to vary six key injection molding process
variables including cavity fill time, plastic melt temperature,
mold surface temperature, cavity holding pressure, cavity
holding time, and mold closed time.  Investigating only two
variables, CMLogic would use nine simulations.  A complete
solution incorporating all six variables requires 45 simulation
jobs and the power of a Cray Research supercomputer is key to
getting those jobs processed fast and profitably, Obermier
said. 

He said that predicting the full process window -- the range of
acceptable settings for all the process variables -- is
essential in setting up the most cost-effective, and productive
plastic injection molding process.  One of CMLogic's unique
features is its ability to let users interactively explore the
results of a CMLogic session, said Obermier.  

"With our CMLogic's molding simulator, users can 'twist knobs'
and try out a mold design while sitting at your workstation,"
he said.  "For example, a user can vary the plastic melt
temperature and see how that affects the pressure needed to
fill the mold -- it really is a tool to provide insight into the
mold and literally see plastic injection molding in process." 

CMLogic is available only on Cray Research supercomputing
systems, which today begin at prices ranging from $150,000 to
$225,000 (U.S.).  The new software is priced according to Cray
Research system and begins at $18,000 for a paid up license. 
 
Cray Research creates the most powerful, highest-quality
computational tools for solving the world's most challenging
scientific and industrial problems.

                               ###