`
`The CRAY-2
`Computer System
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 1
`
`
`
`The CRAY-2 Computer System
`
`t FI
`
`*1I
`
`I
`
`t
`
`*
`
`!
`
`I
`
`-
`
`* <■
`
`11
`
`i—
`s
`
`h
`
`i
`
`i
`»-
`=•
`
`iT
`
`.«
`:S
`
`=
`
`i
`
`Iff
`
`'
`
`r
`
`F
`
`Cray Research,
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 2
`
`
`
`Cray Research's mission is to lead in the development and marketing of
`high-performance systems that make a unique contribution to the markets
`they serve. For close to a decade, Cray Research has been the industry
`leader in large-scale computer systems. Today, the majority of
`supercomputers installed worldwide are Cray systems. These systems are
`used in advanced research laboratories around the world and have gained
`strong acceptance in diverse industrial environments. No other
`manufacturer has Cray Research's breadth of success and experience in
`supercomputer development.
`
`The company's initial product, the CRAY-1 Computer System, was first
`installed in 1976. The CRAY-1 quickly established itself as the standard of
`value for large-scale computers and was soon recognized as the first
`commercially successful vector processor. For some time previously, the
`potential advantages of vector processing had been understood, but
`effective practical implementation had eluded computer architects. The
`CRAY-1 broke that barrier, and today vectorization techniques are used
`commonly by scientists and engineers in a wide variety of disciplines.
`
`With its significant innovations in architecture and technology, the
`CRAY-2 Computer System sets the standard for the next generation of
`supercomputers. The CRAY-2 design allows many types of users to solve
`problems that cannot be solved with any other computers. The CRAY-2
`provides an order of magnitude increase in performance over the CRAY-1
`at an attractive price/performance ratio.
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 3
`
`
`
`The CRAY-2 Computer System sets the standard
`for the next generation of supercomputers. It is
`characterized by a large Common Memory (256
`million 64-bit words), four Background Processors,
`a clock cycle of 4.1 nanoseconds (4.1 billionthsof a
`second) and liquid immersion cooling. It offers
`effective throughput six to twelve times that of the
`CRAY-1 and runs an operating system based on the
`increasingly popular UNIX™ operating system.
`
`The CRAY-2 Computer System uses the most
`advanced technology available. The compact
`mainframe is immersed in a fluorocarbon liquid that
`dissipates the heat generated on the densely
`packed electronic components. The logic and
`memory circuits are contained in eight-layer,
`three-dimensional modules. The large Common
`Memory is constructed of the most dense memory
`chips available, and the logic circuits are
`constructed from the fastest silicon chips available.
`
`The CRAY-2 mainframe contains four independent
`Background Processors, each more powerful than a
`CRAY-1 computer. Featuring a clock cycle time of
`4.1 nanoseconds — faster than any other computer
`system available — each of these processors offers
`exceptional scalar and vector processing
`capabilities. The four Background Processors can
`operate independently on separate jobs or
`concurrently on a single problem. The very
`high-speed Local Memory integral to each
`Background Processor is available for temporary
`storage of vector and scalar data.
`
`Common Memory is one of the most important
`features of the CRAY-2. It consists of 256 million
`64-bit words randomly accessible from any of the
`four Background Processors and from any of the
`high-speed and common data channels. The
`memory is arranged in four quadrants with 128
`interleaved banks. All memory access is performed
`automatically by the hardware. Any user may use all
`or part of this memory.
`
`In conventional memory-limited computer systems,
`I/O wait times for large problems that use
`out-of-memory storage run into hours. With the
`large Common Memory of the CRAY-2, many of
`these problems become CPU-bound.
`
`Introducing the CRAY-2
`Computer System
`
`Cray Research, Inc.
`
`1
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 4
`
`
`
`Control of network access equipment and the
`high-speed disk drives is integral to the CRAY-2
`mainframe hardware. A single Foreground
`Processor coordinates the data flow between the
`system Common Memory and all external devices
`across four high-speed I/O channels. The
`synchronous operation of the Foreground
`Processor with the four Background Processors
`and the external devices provides a significant
`increase in data throughput.
`
`To complement the new CRAY-2 architecture, Cray
`Research has developed an interactive operating
`system based on AT&T's UNIX System V. The
`CRAY-2 Operating System is supported by a
`FORTRAN compiler based on the proven Cray
`Research FORTRAN compiler, CFT.
`
`The CRAY-2 Computer System represents a major
`advance in large-scale computing. The
`combination of four high-speed Background
`Processors, a high-speed Local Memory, a huge
`Common Memory, an extremely powerful I/O
`capability and a comprehensive software product
`offers unsurpassed and balanced performance for
`the user.
`
`Features of the CRAY-2
`• Extremely large directly addressable Common
`Memory
`• Fastest cycle time available in a computer
`system (4.1 nsec)
`• Scalar and vector processing combined with
`multiprocessing
`• Integral Foreground Processor
`• Elegant architecture
`• Extremely high reliability
`• Uses high density memory chips and extremely
`fast silicon logic chips
`• Liquid immersion cooling
`• An operating system based on industry
`recognized UNIX system
`• Automatic vectorizing FORTRAN compiler
`
`2
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 5
`
`
`
`CRAY-2 system overview
`
`Common Memory
`
`Common
`Memory
`Port
`
`Common
`Memory
`Port
`
`Common
`Memory
`Port
`
`Common
`Memory
`Port
`
`Background
`Processor
`
`Background
`Processor
`
`Background
`Processor
`
`Background
`Processor
`
`Disk
`Controllers
`
`Disk
`Controllers
`
`Disk
`Controllers
`
`Disk
`Controllers
`
`Front-end
`Interface
`
`Front-end
`Interface
`
`Front-end
`Interface
`
`Front-end
`Interface
`
`Foreground Processor
`
`Cray Research, Inc.
`
`3
`
`CRAY-2 system overview
`
`Cray Research, Inc.
`
`3
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 6
`
`
`
`Physical characteristics
`
`The CRAY-2 mainframe is elegant in appearance
`as well as in architecture. The memory, computer
`logic and DC power supplies are integrated into a
`compact mainframe composed of 14 vertical
`columns arranged in a 300° arc.
`
`The upper part of each column contains a stack of
`24 modules and the lower part contains power
`supplies for the system. Total cabinet height,
`including the power supplies, is 45 inches, and the
`diameter of the mainframe is 53 inches. Thus, the
`"footprint" of the mainframe is a mere 16 square
`feet of floor space.
`
`An inert fluorocarbon liquid circulates in the
`mainframe cabinet in direct contact with the
`integrated circuit packages. This liquid immersion
`cooling technology allows for the small size of the
`CRAY-2 mainframe and is thus largely responsible
`for the high computation rates.
`
`Physical characteristics
`
`• Occupies only 16 sq ft of floor space
`• Stands 45 inches high; diameter is 53 inches
`• Weighs 5500 pounds
`• 14 columns arranged in a 300° arc
`• Liquid immersion cooling
`• Uses 16-gate array logic chips
`• Three-dimensional modules
`• 320 pluggable modules
`• 195 KW power consumption
`• 400 Hz power from motor generators
`• Chilled water heat exchange
`
`4
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 7
`
`
`
`T
`
`1 '
`
`f-rr
`
`.*;•
`S'-
`
`Cray Research, Inc.
`
`5
`
`Cray Research, Inc.
`
`5
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 8
`
`
`
`Architecture and design
`
`In addition to the cooling technology, the CRAY-2's
`extremely high processing rates are achieved by a
`balanced integration of scalar and vector
`capabilities and a large Common Memory in a
`multiprocessing environment.
`
`The significant architectural components of the
`CRAY-2 Computer System include four identical
`Background Processors, 256 million 64-bit words
`of Common Memory, a Foreground Processor and a
`maintenance control console.
`
`Each of the four identical Background Processors
`contains registers and functional units to perform
`both vector and scalar operations. The single
`Foreground Processor supervises the four
`Background Processors, while the large Common
`Memory complements the processors and provides
`architectural balance, thus assuring extremely high
`throughput rates.
`
`Onsite maintenance is possible via the
`maintenance control console.
`
`Background Processors
`
`Each Background Processor consists of a
`computation section, a control section and a
`high-speed Local Memory. The computation
`section performs arithmetic and logical
`calculations. These operations and the other
`functions of a Background Processor are
`coordinated through the control section. Local
`Memory is used to store temporarily scalar and
`vector data during computations. Each Local
`Memory is 16,384 64-bit words.
`
`Control and data paths for one Background
`Processor are shown in the block diagram
`(opposite page).
`
`Computation section
`The computation section contains registers and
`functional units that operate together to execute a
`program of instructions stored in memory.
`
`Computation section characteristics
`• Twos complement integer and signed
`magnitude floating-point arithmetic
`• Address and arithmetic registers
`- Eight 32-bit address (A) registers
`- Eight 64-bit scalar (S) registers
`- Eight 64-element vector (V) registers; 64 bits
`per element
`• Address functional units
`-Add/subtract
`- Multiply
`• Scalar functional units
`-Add/subtract
`- Shift
`- Logical
`- Population/parity
`-Leading zero count
`• Vector functional units
`- Logical
`- Integer
`Shift
`Add/subtract
`Population/parity
`Leading zero count
`Compressed iota
`• Floating-point functional units
`- Add/subtract
`- Multiply/reciprocal/square root
`• Scatter and gather vector operations to and from
`Common Memory
`
`6
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 9
`
`
`
`CRAY-2 system organization
`
`I
`
`»a>itV,t2.K)TAI Ihr 1s-*—rt
`
`T
`
`AM
`
`’
`
`■
`
`l *,c I
`
`AJ
`
`/.
`V*ctO« lonlfOl M-
`
`rn
`rt
`— j
`
`■B"
`J-CD op q ■I
`
`CK> A D
`
`O.l
`
`>3
`
`■^.yo^naCPUA
`
`.
`
`10
`
`r.
`I [
`
`fORCOAOOMO
`
`Cray Research, Inc.
`
`7
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 10
`
`
`
`Local Memory
`Each Background Processor contains 16,384
`64-bit words of Local Memory. Local Memory is
`treated as a register file to hold scalar operands
`during computation. It may also be used for
`temporary storage of vector segments where these
`segments are used more than once in a
`computation in the vector registers. The access
`time for Local Memory is four clock periods, and
`accesses can overlap accesses to Common
`Memory. This Local Memory replaces the B and T
`registers on the CRAY-1 and is readily available for
`user jobs. One application is for small matrices.
`
`Local Memory characteristics
`
`• 16,384 64-bit words
`• Holds scalar and vector operands during
`computation
`• Temporary storage of vector segments
`• Four-clock-period access time to first word
`• Overlapping register accesses with Common
`Memory accesses
`• Replaces CRAY-1 Band T registers
`
`Control section
`Each Background Processor contains an identical
`independent control section of registers and
`instruction buffers for instruction issue and control.
`
`Control section characteristics
`• Eight instruction buffers, each holding 64 16-bit
`instruction parcels
`• 128 basic instruction codes
`• 32-bit Program Address register
`• 32-bit Base Address register
`• 32-bit Limit Address register
`• 64-bit real-time clock
`• Eight Semaphore flags to provide interlocks for
`Common Memory access
`• 32-bit Status register
`
`Each Background Processor has a 64-bit real-time
`clock. These clocks and the Foreground Processor
`real-time clock are synchronized at system start-up
`and are advanced by one count in each clock
`period.
`
`Background Processor intercommunication
`Synchronization of two or more Background
`Processors cooperating on a single job is achieved
`through use of one of the eight Semaphore flags
`shared by the Background Processors. These flags
`are one-bit registers providing interlocks for
`common access to shared memory fields. A
`Background Processor is assigned access to one
`Semaphore flag by a field in the Status register. The
`Background Processor has instructions to test and
`branch, set and clear a Semaphore flag.
`
`8
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 11
`
`
`
`Foreground Processor and I/O section
`
`The Foreground Processor supervises overall
`system activity among the Foreground Processor,
`Background Processors, Common Memory and
`peripheral controllers. System communication
`occurs through four high-speed synchronous data
`channels.
`
`Firmware control programs for normal system
`operation and a set of diagnostic routines for
`system maintenance are integral to the Foreground
`Processor.
`
`Control circuitry for external devices is also located
`within the CRAY-2 mainframe.
`
`Foreground communication channels
`
`The Foreground Processor is connected to four
`4-gigabit communication channels. These
`channels link the Background Processors,
`Foreground Processor, peripheral controllers and
`Common Memory. Each channel connects one
`Background Processor, one group of peripheral
`controllers, one Common Memory port and the
`Foreground Processor. Data traffic travels directly
`between controllers and Common Memory.
`
`Common Memory
`
`One of the primary technological advantages of the
`CRAY-2 Computer System is its extremely large
`directly addressable Common Memory. Featuring
`268,435,456 words, this Common Memory is
`significantly larger than that offered on any other
`commercially available computer system. It allows
`the individual user to run programs that would be
`impossible to run on any other system. It also
`enhances multiprogramming by allowing an
`exponential Increase in the number of jobs that can
`reside concurrently in memory (that is, that can be
`multiprogrammed).
`
`Common Memory is arranged in four quadrants of
`32 banks each, for a total of 128 banks. A word of
`memory consists of 64 data bits and 8 error
`correction bits (SECDED). This memory is shared
`by the Foreground Processor, Background
`Processors and peripheral equipment controllers.
`Each bank of memory has an independent data
`path to each of the four Common Memory ports.
`Each bi-directional Common Memory port
`connects to a Background Processor and a
`foreground communications channel.
`
`Total memory bandwidth is 64 gigabits or 1 billion
`words per second.
`
`Common Memory characteristics
`
`• 256 million words
`• 64 data bits, 8 error correction bits per word
`• 128 banks; 2 million words per bank
`• Dynamic MOS memory technology
`
`Cray Research, Inc.
`
`9
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 12
`
`
`
`CRAY-2 technology
`
`10
`
`Technological innovations on the CRAY-2 include
`the use of liquid immersion cooling and the
`eight-layer, three-dimensional modules.
`
`Liquid immersion cooling
`
`Effective cooling techniques are central to the
`design of high-speed computational systems.
`Densely packed components result in shorter
`signal paths, thus contributing to higher speeds.
`Traditionally, the tradeoff has been lower reliability
`due to increased operating temperatures, but this is
`no longer a limitation. The liquid immersion cooling
`technology used by the CRAY-2 is a breakthrough
`in the design of cooling systems for large-scale
`computers. It places the cooling medium in direct
`contact with the components to be cooled, thus
`efficiently reducing and stabilizing the operating
`temperature and increasing system reliability.
`
`The CRAY-2 mainframe operates in a cabinet filled
`with a colorless, odorless, inert fluorocarbon fluid.
`The fluid is nontoxic and nonflammable, and has
`high dielectric (insulating) properties. It also has
`high thermal stability and outstanding heat transfer
`properties. The coolant flows through the module
`circuit boards at a velocity of one inch per second
`and is in direct contact with the integrated circuit
`packages and power supplies.
`
`Liquid immersion cooling characteristics
`• The key to densely packed electronics
`• "Valveless" cooling system
`- 200 gallon closed system
`- Room temperature cooling ranges
`- Accompanying stand pipe and reservoir
`- Shell and tube heat exchange
`• Chilled water cooling
`• Fluorocarbon fluid
`- Colorless
`- Odorless
`- Inert: nontoxic and nonflammable
`- High insulant properties
`- High thermal stability
`- High heat transfer capacity
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 13
`
`
`
`I
`
`*
`e:*H
`
`I t if
`
`Lihr-ll
`
`& m
`
`Cray Research, Inc.
`
`11
`
`Cray Research, Inc.
`
`11
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 14
`
`
`
`Module technology design
`
`The CRAY-2 hardware is constructed of
`synchronous networks of binary circuits. These
`circuits are packaged in 320 pluggable modules,
`each of which contains approximately 750
`integrated circuit packages. Total integrated circuit
`population in the system is approximately 240,000
`chips, nearly 75,000 of which are memory.
`
`The pluggable modules are three-dimensional
`structures with an 8 x 8 x 12 array of circuit
`packages. Eight printed circuit boards form the
`module structure. Circuit interconnections are
`made in all three dimensions within the module.
`Each module measures 1x4x8 inches, weighs 2
`pounds, consists of approximately 40% integrated
`circuits by volume and consumes 300 to 500 watts
`of power.
`
`The CRAY-2 Common Memory consists of 128
`memory banks with two million words per bank.
`Each memory bank occupies a circuit module.
`
`CRAY-2 logic networks are constructed of 16-gate
`array integrated circuits packaged in
`three-dimensional structures.
`
`12
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 15
`
`
`
`CRAY-2 reliability
`
`A notable increase in reliability is another benefit of
`the immersion cooling technology. All components
`rapidly dissipate heat to the fluid, thus preventing
`high chip temperatures. These chip temperatures
`are substantially lower than those achieved by other
`types of cooling and result in significantly reduced
`chip failure rates. Efficient heat dissipation also
`prevents destructive thermal shocks that might
`result from large temperature differentials and
`fluctuations.
`
`In addition, a fifteen-to-one decrease in module
`count per CPU from the CRAY-1 and a ten-to-one
`reduction in memory module count enhance failure
`isolation, producing a corresponding increase in
`maintenance efficiency.
`
`CRAY-2 maintenance
`
`If a module should fail, effective and timely
`maintenance is a routine operation. Diagnostic
`software quickly isolates the problem to the failing
`module. The immersion fluid is quickly pumped into
`the reservoir adjacent to the mainframe. The front
`panel is easily removed for ready access to the
`module, which can then be replaced. The front
`panel is then reinstalled and the fluid quickly
`returned to the mainframe. The entire operation
`requires only a few minutes. Once the system is
`restarted, further diagnosis and repair of the faulty
`module can occur on site.
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 16
`
`
`
`CRAY-2 software
`
`Cray Research has made a major commitment to
`the development of a comprehensive and useful
`user environment through an aggressive software
`development program.
`
`The CRAY-2 Computer System comes with
`state-of-the-art software including an operating
`system based on AT&T UNIX System V, an
`automatic vectorizing FORTRAN compiler, a
`comprehensive set of utilities and libraries and a C
`language compiler. The software has extensive
`development objectives beyond initial deliveries,
`including expanded networking capabilities and
`application program migration.
`
`The choice of an operating system based on UNIX
`provides the CRAY-2 user with a well-defined
`program development environment joined with the
`advanced computational power of the CRAY-2. The
`user can access the power of the system through
`the FORTRAN compiler (CFT) and the optimizing
`library routines. CFT is a proven compiler that
`performs automatic vectorization and will conform
`to the ANSI X3.9-1978 FORTRAN 77 standard.
`
`Software summary
`
`• The CRAY-2 Operating System, based on the
`proven UNIX System V and enhanced to fit the
`large-scale scientific computer environment
`• CFT, a vectorizing and optimizing FORTRAN
`compiler
`• An optimized FORTRAN mathematical and I/O
`subroutine library
`• A scientific subroutine library optimized for the
`CRAY-2
`• A multitasking library that allows user
`partitioning of an application into concurrently
`executing tasks
`• A wide variety of system utilities to support the
`needs of interactive and batch processing
`• AC language compiler that supports the needs
`of system software written in C language
`• CAL, the CRAY macro assembler, that provides
`access to all CRAY-2 instructions
`
`14
`
`CRAY-2 Operating System
`
`The CRAY-2 Operating System is based on UNIX
`System V, an operating system developed by AT&T
`Bell Laboratories. In recent years, versions of UNIX
`have become available on many different computer
`systems. UNIX is written in a high-level language
`called C and contains a kernel and a large, diverse
`set of utilities and library programs.
`
`The kernel is the heart of the system. It has a simple,
`well-constructed and clean structure with short and
`efficient software control paths. It supports a small
`number of system call primitives that library and
`application programs can use together to perform
`more complex tasks. The kernel is
`procedure-oriented, encompassing many
`processes that dynamically share a common data
`area used to control the operation of the CRAY-2
`system. The system is oriented towards an
`interactive environment with a hierarchical file
`structure. This structure features directories, user
`ownership and file protection/privacy.
`
`The kernel of the CRAY-2 Operating System has
`been substantially enhanced in the areas of I/O
`processing and in the efficient use of very large
`data files. Other significant enhancements include
`support for asynchronous I/O, improved file system
`reliability, multiprocessing and user multitasking.
`
`Users may initiate asynchronous processes to
`communicate with one another and to pass data
`between them. A variety of command structures
`(shells) are possible. The CRAY-2 Operating
`System offers a standard shell; others may be
`created to provide different command interfaces for
`the users. A batch processing capability is provided
`for efficient use of the system by large, long-running
`jobs.
`
`The operating system supports high-level
`languages (including FORTRAN and C) and the
`mechanism to deliver a common operating system
`environment across a variety of interconnected
`computer systems. It delivers the ultimate in
`computational performance possible on the
`CRAY-2.
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 17
`
`
`
`CRAY-2 FORTRAN Compiler and
`Libraries
`
`The CRAY-2 FORTRAN compiler, CFT Version 2, is
`based on CFT, the highly successful CRAY-1
`compiler that was the first in the industry to
`automatically vectorize codes.
`
`CFT Version 2 automatically vectorizes inner
`DO-loops, provides normal program optimization
`and exploits many of the unique features of the
`CRAY-2 architecture. It does this without sacrificing
`high compilation rates.
`
`The compiler and FORTRAN library offer current
`Cray customers a high level of source code
`compatibility by making available FORTRAN
`extensions, compiler directives and library
`interfaces available on other Cray Research
`products.
`
`The FORTRAN library and a library of highly
`optimized scientific subroutines enable the user to
`take maximum advantage of the architecture of the
`hardware. The I/O library provides the FORTRAN
`user with convenient and efficient use of external
`devices at maximum data rates for large files.
`
`Multitasking
`
`Multitasking is a feature that allows two or more
`parts of a program to be executed in parallel. This
`results in substantial throughput improvements
`over serially executed programs. The performance
`improvements are in proportion to the number of
`tasks that can be constructed for the program and
`the number of Background Processors that can be
`applied to these separate tasks.
`
`In conjunction with vectorization and large memory
`support, a flexible multitasking capability provides
`a major performance step in large-scale scientific
`computing. The user interface to the CRAY-2
`multitasking capability is a set of
`FORTRAN-callable library routines that are
`compatible with similar routines available on other
`Cray products.
`
`CRAY-2 FORTRAN features
`
`• ANSI standard compiler
`• Automatic optimization of code
`- Vectorizes inner loops
`- Uses Local Memory
`• Portability of application codes is a primary goal
`• Very high compilation rates
`• Library routines
`- Scientific library
`- I/O library
`- Multitasking library
`
`C Language
`
`The C programming language is a high-level
`language used extensively in the creation of the
`CRAY-2 Operating System and the majority of the
`utility programs that comprise the system. It is a
`modern computer language that is available on
`processors ranging from microcomputers to
`mainframe computers and now to Cray computers.
`C is useful for a wide range of applications and
`system-oriented programs. The availability of C
`complements the scientific orientation of FORTRAN.
`
`Utilities
`
`A useful and appropriate set of software tools assist
`both interactive and batch users in the efficient use
`of the system. Operational support facilities enable
`proper management of the system.
`
`CAL
`
`The CRAY-2 Assembler, CAL Version 2, provides a
`powerful macro assembly language that allows the
`user to take advantage of all CRAY-2 instructions,
`while using an instruction syntax and macro
`capability that is similar to the CRAY-1 assembler.
`
`Cray Research, Inc.
`
`15
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 18
`
`
`
`Applications
`• Fluid dynamics
`• Circuit simulation and design
`• Structural analysis
`• Energy research
`• Weather forecasting
`• Atmospheric and oceanic research
`• Quantum chemistry
`• Artificial intelligence
`• Genetic engineering
`• Signal image processing
`• Molecular dynamics
`• Petroleum exploration and extraction
`• Process design
`• Economic modeling
`
`Applications
`
`The CRAY-2 Computer System provides balanced
`performance for computationally intensive
`large-scale applications. Generating solutions to
`many important problem classes depends heavily
`on the number of data points that can be
`considered and the number of computations that
`can be performed. The CRAY-2 provides
`substantial increases over its predecessors with
`respect both to the number of data points and the
`computation rate. Researchers and engineers
`realistically can apply the CRAY-2 to problems
`previously considered computationally intractable,
`as well as solving more commonplace problems
`faster and with greater accuracy.
`
`One such application is the simulation of physical
`phenomena — the analysis and prediction of the
`behavior of physical systems through computer
`modeling. Such simulation is common in weather
`forecasting, aircraft and automotive design, energy
`research, geophysical research and seismic
`analysis. The CRAY-2 opens the door to true
`three-dimensional simulation in a wide variety of
`problem domains. The CRAY-2 also offers a
`challenging opportunity for new solutions to
`applications in such fields as genetic engineering,
`artificial intelligence, quantum chemistry and
`economic modeling.
`
`The CRAY-2 offers dramatic improvements in
`throughput via the balanced exploitation of large
`memory, fast vector and scalar computation rates
`and multiprocessing. Problems with previously
`prohibitive I/O requirements can now fit in memory.
`Vectorization and multiprocessing promote very
`high computation rates. In practical terms, this
`means that problems previously considered
`large-scale become medium- or even small-scale
`on the CRAY-2. And problems previously
`considered unsolvable or too costly to solve
`become solvable and economically feasible with
`the CRAY-2.
`
`16
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 19
`
`
`
`Offering advanced architecture, advanced technology and advanced
`software, the CRAY-2 clearly leads the industry in large-scale computing.
`The CRAY-2 leads in technology by offering the fastest processor clock
`cycle (4.1 nanoseconds), the largest memory (256 million words) and four
`vector and scalar multiprocessing Background Processors. The CRAY-2
`leads the industry in computer architecture by applying the fastest or most
`dense components available and packaging them in three-dimensional
`modules immersed in liquid coolant. The CRAY-2 leads the industry in
`software by converting an industry recognized and accepted operating
`system and tailoring it to the needs of large-scale computers, by providing
`a FORTRAN compiler that automatically takes advantage of the system
`architecture and by offering extensions available to the operating systems
`that promote efficient use of the system.
`
`About Cray Research, Inc.
`
`Cray Research, Inc. was organized in 1972 by Seymour R. Cray, a leading
`designer of large-scale scientific computers, and by a small group of
`associates experienced in the computer industry. The company was
`formed to design, develop, manufacture and market large-capacity,
`high-speed computers. The first model produced was the CRAY-1
`Computer System.
`
`Mr. Cray has been a leading architect of large scientific computers for
`more than 25 years. From 1957 to 1968, he served as a director of Control
`Data Corporation (CDC) and was a senior vice president at the time of his
`resignation in early 1972. In that time, he was the principal architect in the
`design and development of the CDC 1604, 6600 and 7600 Computer
`Systems. Prior to his association with CDC, Mr. Cray was employed at the
`Univac Division of Sperry Rand Corporation and its predecessor
`companies, Engineering Research Associates and Remington Rand. Mr.
`Cray has been the principal designer and developer of both the CRAY-1
`and the CRAY-2 Computer Systems.
`
`Today, Cray Research is the world leader in supercomputers, with over
`100 CRAY-1 and CRAY X-MP systems installed worldwide. The company
`employs nearly 2500 people and operates manufacturing, research,
`development and administrative facilities in Chippewa Falls, Wisconsin
`and the Minneapolis, Minnesota area. The company has sixteen domestic
`sales and support offices and eight subsidiary operations in Western
`Europe, Canada and Japan.
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 20
`
`
`
`Corporate Headquarters
`608 Second Avenue South
`Minneapolis, MN 55402
`612/333-5889
`
`MP-0201
`® 1985, Cray Research, Inc.
`
`Domestic sales offices
`
`International subsidiaries
`
`Albuquerque, New Mexico
`Atlanta, Georgia
`Beltsville, Maryland
`Boston, Massachusetts
`Boulder, Colorado
`Chicago, Illinois
`Dallas, Texas
`Dearborn, Michigan
`Houston, Texas
`Laurel, Maryland
`Los Angeles, California
`Minneapolis, Minnesota
`Pittsburgh, Pennsylvania
`Pleasanton, California
`Seattle, Washington
`Tampa, Florida
`Tulsa, Oklahoma
`
`Cray Canada Inc.
`Toronto, Canada
`
`Cray Research France, S.A.
`Paris, France
`
`Cray Research GmbH
`Munich, West Germany
`
`Cray Research Japan, Limited
`Tokyo, Japan
`
`Cray Research (UK) Limited
`Bracknell, Berkshire, UK
`
`The equipment specifications contained in this brochure and the availability of said equipment are subject to change without notice.
`For the latest information, contact your local Cray Research sales office.
`
`CRAY-1 is a registered trademark, and CRAY X-MP and CRAY-2 are trademarks of Cray Research, Inc.
`
`UNIX is a trademark of AT&T Bell Laboratories.
`
`PATENT OWNER DIRECTSTREAM, LLC
`EX. 2164, p. 21
`
`