A supercomputer is a computer that is at the frontline of current processing capacity,

particularly speed of calculation. Supercomputers were introduced in the 1960s and were
designed primarily by Seymour Cray at Control Data Corporation (CDC), which led the
market into the 1970s until Cray left to form his own company, Cray Research. He then
took over the supercomputer market with his new designs, holding the top spot in
supercomputing for five years (1985–1990). In the 1980s a large number of smaller
competitors entered the market, in parallel to the creation of the minicomputer market a
decade earlier, but many of these disappeared in the mid-1990s "supercomputer market
crash".

Today, supercomputers are typically one-of-a-kind custom designs produced by traditional

companies such as Cray, IBM and Hewlett-Packard, who had purchased many of the 1980s
companies to gain their experience. Since October 2010, the Tianhe-1A supercomputer has
been the fastest in the world; it is located in China.

The term supercomputer itself is rather fluid, and the speed of today's supercomputers
tends to become typical of tomorrow's ordinary computers. CDC's early machines were
simply very fast scalar processors, some ten times the speed of the fastest machines offered
by other companies. In the 1970s most supercomputers were dedicated to running a vector
processor, and many of the newer players developed their own such processors at a lower
price to enter the market. The early and mid-1980s saw machines with a modest number of
vector processors working in parallel to become the standard. Typical numbers of
processors were in the range of four to sixteen. In the later 1980s and 1990s, attention
turned from vector processors to massive parallel processing systems with thousands of
"ordinary" CPUs, some being off the shelf units and others being custom designs. Today,
parallel designs are based on "off the shelf" server-class microprocessors, such as the
PowerPC, Opteron, or Xeon, and coprocessors like NVIDIA Tesla GPGPUs, AMD GPUs,
IBM Cell, FPGAs. Most[which?] modern supercomputers are now highly-tuned computer
clusters using commodity processors combined with custom interconnects.

Supercomputers are used for highly calculation-intensive tasks such as problems involving
quantum physics, weather forecasting, climate research, molecular modeling (computing
the structures and properties of chemical compounds, biological macromolecules,
polymers, and crystals), physical simulations (such as simulation of airplanes in wind
tunnels, simulation of the detonation of nuclear weapons, and research into nuclear fusion).

Relevant here is the distinction between capability computing and capacity computing, as
defined by Graham et al. Capability computing is typically thought of as using the
maximum computing power to solve a large problem in the shortest amount of time. Often
a capability system is able to solve a problem of a size or complexity that no other
computer can. Capacity computing in contrast is typically thought of as using efficient cost-
effective computing power to solve somewhat large problems or many small problems or to
prepare for a run on a capability system.

[edit] Hardware and software design

 This section does not cite any references or sources.
Please help improve this article by adding citations to reliable sources. Unsourced material may be
challenged and removed. (July 2008)

Processor board of a CRAY YMP vector computer

Supercomputers using custom CPUs traditionally gained their speed over conventional
computers through the use of innovative designs that allow them to perform many tasks in
parallel, as well as complex detail engineering. They tend to be specialized for certain types
of computation, usually numerical calculations, and perform poorly at more general
computing tasks. Their memory hierarchy is very carefully designed to ensure the
processor is kept fed with data and instructions at all times — in fact, much of the
performance difference between slower computers and supercomputers is due to the
memory hierarchy. Their I/O systems tend to be designed to support high bandwidth, with
latency less of an issue, because supercomputers are not used for transaction processing.

As with all highly parallel systems, Amdahl's law applies, and supercomputer designs
devote great effort to eliminating software serialization, and using hardware to address the
remaining bottlenecks.

[edit] Supercomputer challenges, technologies

 A supercomputer consumes large amounts of electrical power, almost all of which is

converted into heat, requiring cooling. For example, Tianhe-1A consumes 4.04
Megawatts of electricity.[1] The cost to power and cool the system is usually one of
the factors that limit the scalability of the system. (For example, 4MW at
$0.10/KWh is $400 an hour or about $3.5 million per year).
 Information cannot move faster than the speed of light between two parts of a
supercomputer. For this reason, a supercomputer that is many meters across must
have latencies between its components measured at least in the tens of nanoseconds.
Seymour Cray's supercomputer designs attempted to keep cable runs as short as
possible for this reason, hence the cylindrical shape of his Cray range of computers.
In modern supercomputers built of many conventional CPUs running in parallel,
latencies of 1–5 microseconds to send a message between CPUs are typical.
 Supercomputers consume and produce massive amounts of data in a very short
period of time. According to Ken Batcher, "A supercomputer is a device for turning
compute-bound problems into I/O-bound problems." Much work on external
storage bandwidth is needed to ensure that this information can be transferred
quickly and stored/retrieved correctly.
Technologies developed for supercomputers include:

 Vector processing
 Liquid cooling
 Non-Uniform Memory Access (NUMA)
 Striped disks (the first instance of what was later called RAID)
 Parallel filesystems

[edit] Processing techniques

Vector processing techniques were first developed for supercomputers and continue to be
used in specialist high-performance applications. Vector processing techniques have
trickled down to the mass market in DSP architectures and SIMD (Single Instruction
Multiple Data) processing instructions for general-purpose computers.

Modern video game consoles in particular use SIMD extensively and this is the basis for
some manufacturers' claim that their game machines are themselves supercomputers.
Indeed, some graphics cards have the computing power of several TeraFLOPS. The
applications to which this power can be applied was limited by the special-purpose nature
of early video processing. As video processing has become more sophisticated, graphics
processing units (GPUs) have evolved to become more useful as general-purpose vector
processors, and an entire computer science sub-discipline has arisen to exploit this
capability: General-Purpose Computing on Graphics Processing Units (GPGPU).

The current Top500 list (from May 2010) has 3 supercomputers based on GPGPUs. In
particular, the number 3 supercomputer, Nebulae built by Dawning in China, is based on
GPGPUs.[2]

[edit] Operating systems

More than 90% of today's Supercomputers run some variant of Linux.[3]

Supercomputers today most often use variants of Linux.[3] as shown by the graph to the
right.

Until the early-to-mid-1980s, supercomputers usually sacrificed instruction set

compatibility and code portability for performance (processing and memory access speed).
For the most part, supercomputers to this time (unlike high-end mainframes) had vastly
different operating systems. The Cray-1 alone had at least six different proprietary OSs
largely unknown to the general computing community. In a similar manner, different and
incompatible vectorizing and parallelizing compilers for Fortran existed. This trend would
have continued with the ETA-10 were it not for the initial instruction set compatibility
between the Cray-1 and the Cray X-MP, and the adoption of computer systems such as
Cray's Unicos, or Linux.

[edit] Programming

The parallel architectures of supercomputers often dictate the use of special programming
techniques to exploit their speed. The base language of supercomputer code is, in general,
Fortran or C, using special libraries to share data between nodes. In the most common
scenario, environments such as PVM and MPI for loosely connected clusters and OpenMP
for tightly coordinated shared memory machines are used. Significant effort is required to
optimize a problem for the interconnect characteristics of the machine it will be run on; the
aim is to prevent any of the CPUs from wasting time waiting on data from other nodes. The
new massively parallel GPGPUs have hundreds of processor cores and are programmed
using programming models such as CUDA and OpenCL.

[edit] Software tools

Software tools for distributed processing include standard APIs such as MPI and PVM,
VTL, and open source-based software solutions such as Beowulf, WareWulf, and
openMosix, which facilitate the creation of a supercomputer from a collection of ordinary
workstations or servers. Technology like ZeroConf (Rendezvous/Bonjour) can be used to
create ad hoc computer clusters for specialized software such as Apple's Shake compositing
application. An easy programming language for supercomputers remains an open research
topic in computer science. Several utilities that would once have cost several thousands of
dollars are now completely free thanks to the open source community that often creates
disruptive technology[citation needed].

[edit] Modern supercomputer architecture

‹ The template below (Ref improve section) is being considered for deletion. See templates for discussion to help reach a consensus.›
This section needs additional citations for verification.
Please help improve this article by adding reliable references. Unsourced material may be
challenged and removed. (October 2010)
IBM Roadrunner - LANL

The CPU Architecture Share of Top500 Rankings between 1993 and 2009.

Supercomputers today often have a similar top-level architecture consisting of a cluster of

MIMD multiprocessors, each processor of which is SIMD. The supercomputers vary
radically with respect to the number of multiprocessors per cluster, the number of
processors per multiprocessor, and the number of simultaneous instructions per SIMD
processor. Within this hierarchy we have:

 A computer cluster is a collection of computers that are highly interconnected via a

high-speed network or switching fabric. Each computer runs under a separate
instance of an Operating System (OS).
 A multiprocessing computer is a computer, operating under a single OS and using
more than one CPU, wherein the application-level software is indifferent to the
number of processors. The processors share tasks using Symmetric multiprocessing
(SMP) and Non-Uniform Memory Access (NUMA).
 A SIMD processor executes the same instruction on more than one set of data at the
same time. The processor could be a general purpose commodity processor or
special-purpose vector processor. It could also be high-performance processor or a
low power processor. As of 2007, the processor executes several SIMD instructions
per nanosecond.

As of October 2010 the fastest supercomputer in the world is the Tianhe-1A system at
National University of Defense Technology with more than 21000 computers, it boasts a
speed of 2.507 petaflops, over 30% faster than the world's next fastest computer, the Cray
XT5 "Jaguar".

In February 2009, IBM also announced work on "Sequoia," which appears to be a 20

petaflops supercomputer. This will be equivalent to 2 million laptops (whereas Roadrunner
is comparable to a mere 100,000 laptops). It is slated for deployment in late 2011. [4] The
Sequoia will be powered by 1.6 million cores (specific 45-nanometer chips in development)
and 1.6 petabytes of memory. It will be housed in 96 refrigerators spanning roughly
3,000 square feet (280 m2).[5]

Moore's Law and economies of scale are the dominant factors in supercomputer design.
The design concepts that allowed past supercomputers to out-perform desktop machines of
the time tended to be gradually incorporated into commodity PCs. Furthermore, the costs
of chip development and production make it uneconomical to design custom chips for a
small run and favor mass-produced chips that have enough demand to recoup the cost of
production. A current model quad-core Xeon workstation running at 2.66 GHz will
outperform a multimillion dollar Cray C90 supercomputer used in the early 1990s; most
workloads requiring such a supercomputer in the 1990s can be done on workstations
costing less than 4,000 US dollars as of 2010. Supercomputing is taking a step of increasing
density, allowing for desktop supercomputers to become available, offering the computer
power that in 1998 required a large room to require less than a desktop footprint.

In addition, many problems carried out by supercomputers are particularly suitable for
parallelization (in essence, splitting up into smaller parts to be worked on simultaneously)
and, in particular, fairly coarse-grained parallelization that limits the amount of
information that needs to be transferred between independent processing units. For this
reason, traditional supercomputers can be replaced, for many applications, by "clusters"
of computers of standard design, which can be programmed to act as one large computer.

[edit] Special-purpose supercomputers

‹ The template below (Ref improve section) is being considered for deletion. See templates for discussion to help reach a consensus.›
This section needs additional citations for verification.
Please help improve this article by adding reliable references. Unsourced material may be
challenged and removed. (October 2010)

Special-purpose supercomputers are high-performance computing devices with a hardware

architecture dedicated to a single problem. This allows the use of specially programmed
FPGA chips or even custom VLSI chips, allowing higher price/performance ratios by
sacrificing generality. They are used for applications such as astrophysics computation and
brute-force codebreaking. Historically a new special-purpose supercomputer has
occasionally been faster than the world's fastest general-purpose supercomputer, by some
measure. For example, GRAPE-6 was faster than the Earth Simulator in 2002 for a
particular special set of problems.

Examples of special-purpose supercomputers:

  Belle, Deep Blue, and Hydra, for playing chess
 Reconfigurable computing machines or parts of machines
 GRAPE, for astrophysics and molecular dynamics
 Deep Crack, for breaking the DES cipher
 MDGRAPE-3, for protein structure computation
 D. E. Shaw Research Anton, for simulating molecular dynamics [6]
 QPACE, for simulations of the strong interaction (Lattice QCD)

[edit] The fastest supercomputers today

[edit] Measuring supercomputer speed

14 countries account for the vast majority of the world's 500 fastest supercomputers, with
over half being located in the United States.

In general, the speed of a supercomputer is measured in "FLOPS" (FLoating Point

Operations Per Second), commonly used with an SI prefix such as tera-, combined into the
shorthand "TFLOPS" (1012 FLOPS, pronounced teraflops), or peta-, combined into the
shorthand "PFLOPS" (1015 FLOPS, pronounced petaflops.) This measurement is based on
a particular benchmark, which does LU decomposition of a large matrix. This mimics a
class of real-world problems, but is significantly easier to compute than a majority of
actual real-world problems.

"Petascale" supercomputers can process one quadrillion (10 15) (1000 trillion) FLOPS.
Exascale is computing performance in the exaflops range. An exaflop is one quintillion
(1018) FLOPS (one million teraflops).

[edit] The TOP500 list

Main article: TOP500

Since 1993, the fastest supercomputers have been ranked on the TOP500 list according to
their LINPACK benchmark results. The list does not claim to be unbiased or definitive,
but it is a widely cited current definition of the "fastest" supercomputer available at any
given time.
[edit] Current fastest supercomputer system

Tianhe-1A is ranked on the TOP500 list as the fastest supercomputer. It consists of 14,336
Intel Xeon CPUs and 7,168 Nvidia Tesla M2050 GPUs with a new interconnect fabric of
Chinese origin, reportedly twice the speed of InfiniBand.[7] Tianhe-1A spans 103 cabinets,
weighs 155 tons, and consumes 4.04 megawatts of electricity.[8]

A Brief History of the Computer (b.c. – 1993a.d.)

2tweetsretweet

by Jeremy Meyers Note: Yes, a lot of this is from Grolier’s Encyclopedia. Hey, I was
young. I didn’t know any better. Credit where credit is due. Also, this information is only
current as of the early 1990′s (1993, to be exact), and no I’m not planning to add more
information anytime soon. I would not recommend scamming this for your own homework,
as some of the conclusions are rather humorous today. 0digg
Citing This Work You are welcome to use this document as a
reference in creating your own paper or research work on the subject. Please don’t just copy this
paper verbatim and submit it as your own work, as I put a lot of time and effort into it. Plus, it’s bad
karma. If you would like to use this work, please use this citation in your bibliography: Meyers,
Jeremy, “A Short History of the Computer” [Online] Available ([current date])

Table of Contents:

1) In The Beginning… 5) The Modern “Stored Program” EDC

2) Babbage 6) Advances in the 1950s

3) Use of Punched Cards by Hollerith 7) Advances in the 1960s

4) Electronic Digital Computers 8) Recent Advances

In The Beginning…

The history of computers starts out about 2000 years ago, at the birth of the abacus, a
wooden rack holding two horizontal wires with beads strung on them. When these beads
are moved around, according to programming rules memorized by the user, all regular
arithmetic problems can be done. Another important invention around the same time was
the Astrolabe, used for navigation. Blaise Pascal is usually credited for building the first
digital computer in 1642. It added numbers entered with dials and was made to help his
father, a tax collector. In 1671, Gottfried Wilhelm von Leibniz invented a computer that
was built in 1694. It could add, and, after changing some things around, multiply. Leibniz
invented a special stepped gear mechanism for introducing the addend digits, and this is
still being used. The prototypes made by Pascal and Leibniz were not used in many places,
and considered weird until a little more than a century later, when Thomas of Colmar
(A.K.A. Charles Xavier Thomas) created the first successful mechanical calculator that
could add, subtract, multiply, and divide. A lot of improved desktop calculators by many
inventors followed, so that by about 1890, the range of improvements included:

 Accumulation of partial results

 Storage and automatic reentry of past results (A memory function)
 Printing of the results

Each of these required manual installation. These improvements were mainly made for
commercial users, and not for the needs of science.

Babbage
 While Thomas of Colmar was developing the desktop calculator,
a series of very interesting developments in computers was started in Cambridge, England,
by Charles Babbage (left, of which the computer store “Babbages, now GameStop, is
named), a mathematics professor. In 1812, Babbage realized that many long calculations,
especially those needed to make mathematical tables, were really a series of predictable
actions that were constantly repeated. From this he suspected that it should be possible to
do these automatically. He began to design an automatic mechanical calculating machine,
which he called a difference engine. By 1822, he had a working model to demonstrate with.
With financial help from the British government, Babbage started fabrication of a
difference engine in 1823. It was intended to be steam powered and fully automatic,
including the printing of the resulting tables, and commanded by a fixed instruction
program. The difference engine, although having limited adaptability and applicability,
was really a great advance. Babbage continued to work on it for the next 10 years, but in
1833 he lost interest because he thought he had a better idea — the construction of what
would now be called a general purpose, fully program-controlled, automatic mechanical
digital computer. Babbage called this idea an Analytical Engine. The ideas of this design
showed a lot of foresight, although this couldn’t be appreciated until a full century later.
The plans for this engine required an identical decimal computer operating on numbers of
50 decimal digits (or words) and having a storage capacity (memory) of 1,000 such digits.
The built-in operations were supposed to include everything that a modern general –
purpose computer would need, even the all important Conditional Control Transfer
Capability that would allow commands to be executed in any order, not just the order in
which they were programmed. The analytical engine was soon to use punched cards
(similar to those used in a Jacquard loom), which would be read into the machine from
several different Reading Stations. The machine was supposed to operate automatically, by
steam power, and require only one person there. Babbage‘s computers were never finished.
Various reasons are used for his failure. Most used is the lack of precision machining
techniques at the time. Another speculation is that Babbage was working on a solution of a
problem that few people in 1840 really needed to solve. After Babbage, there was a
temporary loss of interest in automatic digital computers. Between 1850 and 1900 great
advances were made in mathematical physics, and it came to be known that most
observable dynamic phenomena can be identified by differential equations(which meant that
most events occurring in nature can be measured or described in one equation or another),
so that easy means for their calculation would be helpful. Moreover, from a practical view,
the availability of steam power caused manufacturing (boilers), transportation (steam
engines and boats), and commerce to prosper and led to a period of a lot of engineering
achievements. The designing of railroads, and the making of steamships, textile mills, and
bridges required differential calculus to determine such things as:

 center of gravity
 center of buoyancy
 moment of inertia
 stress distributions

Even the assessment of the power output of a steam engine needed mathematical
integration. A strong need thus developed for a machine that could rapidly perform many
repetitive calculations.

Use of Punched Cards by Hollerith

A step towards automated computing was the development of punched

cards, which were first successfully used with computers in 1890 by Herman Hollerith (left)
and James Powers, who worked for the US. Census Bureau. They developed devices that
could read the information that had been punched into the cards automatically, without
human help. Because of this, reading errors were reduced dramatically, work flow
increased, and, most importantly, stacks of punched cards could be used as easily
accessible memory of almost unlimited size. Furthermore, different problems could be
stored on different stacks of cards and accessed when needed. These advantages were seen
by commercial companies and soon led to the development of improved punch-card using
computers created by International Business Machines (IBM), Remington (yes, the same
people that make shavers), Burroughs, and other corporations. These computers used
electromechanical devices in which electrical power provided mechanical motion — like
turning the wheels of an adding machine. Such systems included features to:

 feed in a specified number of cards automatically

 add, multiply, and sort
 feed out cards with punched results

As compared to today’s machines, these computers were slow, usually processing 50 – 220
cards per minute, each card holding about 80 decimal numbers (characters). At the time,
however, punched cards were a huge step forward. They provided a means of I/O, and
memory storage on a huge scale. For more than 50 years after their first use, punched card
machines did most of the world’s first business computing, and a considerable amount of
the computing work in science.

Electronic Digital Computers

The start of World War II produced a large need for computer

capacity, especially for the military. New weapons were made for which trajectory tables
and other essential data were needed. In 1942, John P. Eckert, John W. Mauchly (left), and
their associates at the Moore school of Electrical Engineering of University of Pennsylvania
decided to build a high – speed electronic computer to do the job. This machine became
known as ENIAC (Electrical Numerical Integrator And Calculator) The size of ENIAC‘s
numerical “word” was 10 decimal digits, and it could multiply two of these numbers at a
rate of 300 per second, by finding the value of each product from a multiplication table
stored in its memory. ENIAC was therefore about 1,000 times faster then the previous
generation of relay computers. ENIAC used 18,000 vacuum tubes, about 1,800 square feet
of floor space, and consumed about 180,000 watts of electrical power. It had punched card
I/O, 1 multiplier, 1 divider/square rooter, and 20 adders using decimal ring counters, which
served as adders and also as quick-access (.0002 seconds) read-write register storage. The
executable instructions making up a program were embodied in the separate “units” of
ENIAC, which were plugged together to form a “route” for the flow of information.
 These connections had to be redone after each
computation, together with presetting function tables and switches. This “wire your own”
technique was inconvenient (for obvious reasons), and with only some latitude could
ENIAC be considered programmable. It was, however, efficient in handling the particular
programs for which it had been designed. ENIAC is commonly accepted as the first
successful high – speed electronic digital computer (EDC) and was used from 1946 to 1955.
A controversy developed in 1971, however, over the patentability of ENIAC‘s basic digital
concepts, the claim being made that another physicist, John V. Atanasoff (left) had already
used basically the same ideas in a simpler vacuum – tube device he had built in the 1930′s
while at Iowa State College. In 1973 the courts found in favor of the company using the
Atanasoff claim.

The Modern Stored Program EDC

Fascinated by the success of ENIAC, the mathematician John

Von Neumann (left) undertook, in 1945, an abstract study of computation that showed that
a computer should have a very simple, fixed physical structure, and yet be able to execute
any kind of computation by means of a proper programmed control without the need for
any change in the unit itself. Von Neumann contributed a new awareness of how practical,
yet fast computers should be organized and built. These ideas, usually referred to as the
stored – program technique, became essential for future generations of high – speed digital
computers and were universally adopted.
The Stored – Program technique involves many features of computer design and function
besides the one that it is named after. In combination, these features make very – high –
speed operation attainable. A glimpse may be provided by considering what 1,000
operations per second means. If each instruction in a job program were used once in
consecutive order, no human programmer could generate enough instruction to keep the
computer busy. Arrangements must be made, therefore, for parts of the job program
(called subroutines) to be used repeatedly in a manner that depends on the way the
computation goes. Also, it would clearly be helpful if instructions could be changed if
needed during a computation to make them behave differently.

Von Neumann met these two needs by making a special type of machine instruction, called
a Conditional control transfer – which allowed the program sequence to be stopped and
started again at any point – and by storing all instruction programs together with data in
the same memory unit, so that, when needed, instructions could be arithmetically changed
in the same way as data. As a result of these techniques, computing and programming
became much faster, more flexible, and more efficient with work. Regularly used
subroutines did not have to be reprogrammed for each new program, but could be kept in
“libraries” and read into memory only when needed. Thus, much of a given program could
be assembled from the subroutine library.

The all – purpose computer memory became the assembly place in which all parts of a long
computation were kept, worked on piece by piece, and put together to form the final
results. The computer control survived only as an “errand runner” for the overall process.
As soon as the advantage of these techniques became clear, they became a standard
practice.

The first generation of modern programmed electronic computers to take advantage of

these improvements were built in 1947. This group included computers using Random –
Access – Memory (RAM), which is a memory designed to give almost constant access to
any particular piece of information. . These machines had punched – card or punched tape
I/O devices and RAM’s of 1,000 – word capacity and access times of .5 Greek MU seconds
(.5*10-6 seconds). Some of them could perform multiplications in 2 to 4 MU seconds.

Physically, they were much smaller than ENIAC. Some were about the size of a grand
piano and used only 2,500 electron tubes, a lot less then required by the earlier ENIAC.
The first – generation stored – program computers needed a lot of maintenance, reached
probably about 70 to 80% reliability of operation (ROO) and were used for 8 to 12 years.
They were usually programmed in ML, although by the mid 1950′s progress had been
made in several aspects of advanced programming. This group of computers included
EDVAC (above) and UNIVAC (right) the first commercially available computers.

Advances in the 1950′s

Early in the 50′s two important engineering discoveries changed the image of the electronic
– computer field, from one of fast but unreliable hardware to an image of relatively high
reliability and even more capability. These discoveries were the magnetic core memory and
the Transistor – Circuit Element. These technical discoveries quickly found their way into
new models of digital computers. RAM capacities increased from 8,000 to 64,000 words in
commercially available machines by the 1960′s, with access times of 2 to 3 MS
(Milliseconds). These machines were very expensive to purchase or even to rent and were
particularly expensive to operate because of the cost of expanding programming. Such
computers were mostly found in large computer centers operated by industry, government,
and private laboratories – staffed with many programmers and support personnel.

This situation led to modes of operation enabling the sharing of the high potential
available. One such mode is batch processing, in which problems are prepared and then
held ready for computation on a relatively cheap storage medium. Magnetic drums,
magnetic – disk packs, or magnetic tapes were usually used. When the computer finishes
with a problem, it “dumps” the whole problem (program and results) on one of these
peripheral storage units and starts on a new problem. Another mode for fast, powerful
machines is called time-sharing. In time-sharing, the computer processes many jobs in such
rapid succession that each job runs as if the other jobs did not exist, thus keeping each
“customer” satisfied. Such operating modes need elaborate executable programs to attend
to the administration of the various tasks.

Advances in the 1960′s

In the 1960′s, efforts to design and develop the fastest possible computer with the greatest
capacity reached a turning point with the LARC machine, built for the Livermore
Radiation Laboratories of the University of California by the Sperry – Rand Corporation,
and the Stretch computer by IBM. The LARC had a base memory of 98,000 words and
multiplied in 10 Greek MU seconds. Stretch was made with several degrees of memory
having slower access for the ranks of greater capacity, the fastest access time being less
then 1 Greek MU Second and the total capacity in the vicinity of 100,000,000 words.
During this period, the major computer manufacturers began to offer a range of
capabilities and prices, as well as accessories such as:
  Consoles
 Card Feeders
 Page Printers
 Cathode – ray – tube displays
 Graphing devices

These were widely used in businesses for such things as:

 Accounting
 Payroll
 Inventory control
 Ordering Supplies
 Billing

CPU’s for these uses did not have to be very fast arithmetically and were usually used to
access large amounts of records on file, keeping these up to date. By far, the most number
of computer systems were sold for the more simple uses, such as hospitals (keeping track of
patient records, medications, and treatments given). They were also used in libraries, such
as the National Medical Library retrieval system, and in the Chemical Abstracts System,
where computer records on file now cover nearly all known chemical compounds.

More Recent Advances

The trend during the 1970′s was, to some extent, moving away from very powerful, single –
purpose computers and toward a larger range of applications for cheaper computer
systems. Most continuous-process manufacturing, such as petroleum refining and
electrical-power distribution systems, now used computers of smaller capability for
controlling and regulating their jobs.

In the 1960′s, the problems in programming applications were an obstacle to the

independence of medium sized on-site computers, but gains in applications programming
language technologies removed these obstacles. Applications languages were now available
for controlling a great range of manufacturing processes, for using machine tools with
computers, and for many other things. Moreover, a new revolution in computer hardware
was under way, involving shrinking of computer-logic circuitry and of components by what
are called large-scale integration (LSI techniques.

In the 1950s it was realized that “scaling down” the size of electronic digital computer
circuits and parts would increase speed and efficiency and by that, improve performance, if
they could only find a way to do this. About 1960 photo printing of conductive circuit
boards to eliminate wiring became more developed. Then it became possible to build
resistors and capacitors into the circuitry by the same process. In the 1970′s, vacuum
deposition of transistors became the norm, and entire assemblies, with adders, shifting
registers, and counters, became available on tiny “chips.”
In the 1980′s, very large scale integration (VLSI), in which hundreds of thousands of
transistors were placed on a single chip, became more and more common. Many
companies, some new to the computer field, introduced in the 1970s programmable
minicomputers supplied with software packages. The “shrinking” trend continued with the
introduction of personal computers (PC’s), which are programmable machines small
enough and inexpensive enough to be purchased and used by individuals. Many companies,
such as Apple Computer and Radio Shack, introduced very successful PC’s in the 1970s,
encouraged in part by a fad in computer (video) games. In the 1980s some friction occurred
in the crowded PC field, with Apple and IBM keeping strong. In the manufacturing of
semiconductor chips, the Intel and Motorola Corporations were very competitive into the
1980s, although Japanese firms were making strong economic advances, especially in the
area of memory chips.

By the late 1980s, some personal computers were run by microprocessors that, handling 32
bits of data at a time, could process about 4,000,000 instructions per second.
Microprocessors equipped with read-only memory (ROM), which stores constantly used,
unchanging programs, now performed an increased number of process-control, testing,
monitoring, and diagnosing functions, like automobile ignition systems, automobile-engine
diagnosis, and production-line inspection duties. Cray Research and Control Data Inc.
dominated the field of supercomputers, or the most powerful computer systems, through
the 1970s and 1980s.

In the early 1980s, however, the Japanese government announced a gigantic plan to design
and build a new generation of supercomputers. This new generation, the so-called “fifth”
generation, is using new technologies in very large integration, along with new
programming languages, and will be capable of amazing feats in the area of artificial
intelligence, such as voice recognition.

Progress in the area of software has not matched the great advances in hardware. Software
has become the major cost of many systems because programming productivity has not
increased very quickly. New programming techniques, such as object-oriented
programming, have been developed to help relieve this problem. Despite difficulties with
software, however, the cost per calculation of computers is rapidly lessening, and their
convenience and efficiency are expected to increase in the early future. The computer field
continues to experience huge growth. Computer networking, computer mail, and electronic
publishing are just a few of the applications that have grown in recent years. Advances in
technologies continue to produce cheaper and more powerful computers offering the
promise that in the near future, computers or terminals will reside in most, if not all homes,
offices, and schools.