Media:      Steve Conway 612/683-7133 
Financial:  Bill Gacki 612/683-7372

CRAY T3D SYSTEM DEMONSTRATES INDUSTRY-LEADING
PERFORMANCE ON LINPACK BENCHMARK TEST

Top Performance On NAS Parallel Benchmarks Reported Earlier

PORTLAND, Ore., Nov. 16, 1993 -- Cray Research, Inc. (NYSE:CYR) today
reported that the CRAY T3D massively parallel processing (MPP) system,
announced Sept. 27, has achieved industry-leading performance on the highly
parallel version of LINPACK, a widely recognized independent benchmark test for
high-performance computer systems.

On this test, a 128-processor CRAY T3D system achieved a sustained
performance of 9.4 gigaflops (billion floating point operations per second) on
a size 12,000 problem, said Cray Research vice president of technology Steve
Nelson.  By comparison, the next-best performance by a second leading MPP
vendor's 128-processor system on a problem of this size is 4.0 gigaflops.  A
third leading MPP vendor's 128-processor system sustained 7.6 gigaflops, but
required a much larger problem size to attain this speed.  "The ability to
achieve very fast speeds even on smaller problems is a distinct advantage for
users," Nelson said. 

"The CRAY T3D system's industry-leading LINPACK performance is further
proof that this is the most powerful MPP system in the world," he said.  He
noted an earlier report that the CRAY T3D system achieved higher sustained
speeds than any other MPP system on all eight tests of the NAS Parallel
Benchmarks, another widely accepted benchmark suite developed by the NASA Ames
Research Center.  "On some of the NAS Parallel Benchmark tests," he said, "the
CRAY T3D system was as much as four times faster than all other MPP products
with up to 128 processors."

Because of the system's low latency and unrivaled bisection bandwidth (the
amount of data that can be transferred among processors in a given time), the
CRAY T3D's speed advantage over competing systems is expected to increase
substantially in larger system sizes, Nelson said.  "With the T3D, you won't
see the performance-degrading traffic jams' that have plagued other MPP
systems.  The larger the system size compared, the better we will look."

Results for the highly parallel LINPACK benchmark are published periodically
by the test's originator, Jack Dongarra, in a report entitled "Performance of
Various Computers Using Standard Linear Equations Software."  The benchmark
focuses on the solution of dense systems of linear equations, Nelson said.

Cray Research creates the most powerful, highest-quality computational tools
for solving the world's most challenging industrial and scientific problems. 

###