A
address
A character or group of characters that identifies a register, a device, a
particular part of storage, or some other data source or destination.
AIX
Abbreviation for Advanced Interactive Executive, IBM's licensed
version of the UNIX operating system. AIX is particularly suited to support
technical computing applications, including high function graphics and
floating point computations.
Amd
Berkely Software Distribution automount daemon.
APAR
Authorized Program Analysis Report. A report of a problem caused by a
suspected defect in a current unaltered release of a program.
application
The use to which a data processing system is put; for example, a
computational chemistry application, a signal processing application.
application data
The data that is produced using an application program.
argument
A parameter passed between a calling program and a SUBROUTINE subprogram,
a FUNCTION subprogram, or a statement function.
array
An ordered set of data items identified by a single name.
array descriptor
Contains the information required to establish the mapping between a
global data structure and its corresponding process and memory location.
array element
A data item in an array, identified by the array name followed by a
subscript indicating its position in the array.
array name
The name of an ordered set of data items that make up an array.
assignment statement
A statement that assigns a value to a variable or array element. It is
made up of a variable or array element, followed by an equal sign (=),
followed by an expression. The variable, array element, or expression can be
character, logical, or arithmetic. When the assignment statement is processed,
the expression to the right of the equal sign replaces the value of the
variable or array element to the left.
B
bandwidth
The total available bit rate of a digital channel.
Basic Linear Algebra Communication Subprograms (BLACS)
A standard set of public domain subroutines that perform message passing
(communications) between processes.
Basic Linear Algebra Subprograms (BLAS)
A standard set of public domain mathematical subroutines that perform
linear algebra operations.
BLACS
Basic Linear Algebra Communication Subprograms.
BLAS
Basic Linear Algebra Subprograms.
C
cache
A high-speed buffer.
character constant
A string of one or more alphanumeric characters enclosed in apostrophes.
The delimiting apostrophes are not part of the value of the constant.
character expression
An expression in the form of a single character constant, variable, array
element, substring, function reference, or another expression enclosed in
parentheses. A character expression is always of type character.
character type
The data type for representing strings of alphanumeric characters; in
storage, one byte is used for each character.
client
* (1) A function that requests services from a server, and makes them
available to the user. * (2) A term used in an environment to identify a
machine that uses the resources of the network.
cluster
A group of processors interconnected through a high speed network that can
be used for high performance computing.
column-major order
A sequencing method used for storing multidimensional arrays according to
the subscripts of the array elements. In this method the leftmost subscript
position varies most rapidly and completes a full cycle before the next
subscript position to the right is incremented.
CMI
Centralized Management Interface provides a series of SMIT menus and
dialogues used for defining and querying the SP system configuration.
complex conjugate even data
Complex data that has its real part even and its imaginary part odd.
complex constant
An ordered pair of real or integer constants separated by a comma and
enclosed in parentheses. The first real constant of the pair is the real part
of the complex number; the second is the imaginary part.
complex type
The data type for representing an approximation of the value of a complex
number. A data item of this type consists of an ordered pair of real data
items separated by a comma and enclosed in parentheses. The first item
represents the real part of the complex number; the second represents the
imaginary part.
constant
An unvarying quantity. The four classes of constants specify numbers
(arithmetic), truth values (logical), character data (character), and
hexadecimal data.
D
daemon
A process, not associated with a particular user, that performs
system-wide functions such as administration and control of networks,
execution of time-dependent activities, line printer spooling, etc.
default
An alternative value, attribute, or option that is assumed when none has
been specified.
dataless workstation
A workstation that has local disks which may be used for swap,
tmp, and usr file systems.
data distribution
The method in which global data structures are divided among processes.
Three types of data distribution are: cyclic, block-cyclic, and block
distribution.
data type
The structural characteristics, features and properties of data that may
be directly specified by a programming language; for example, integers, real
numbers in Fortran; arrays in APL; linked lists in LISP; character string in
SNOBOL.
decimation
The formation of a sequence containing every n-th element of another
sequence.
dimension of an array
One of the subscript expression positions in a subscript for an array. In
Fortran, an array may have from one to seven dimensions. Graphically, the
first dimension is represented by the rows, the second by the columns, and the
third by the planes. Contrast with rank. See also extent of a dimension.
direct access storage
A storage device in which the access time is in effect independent of the
location of the data. (A)
diskless workstation
A computer workstation with its own processor, keyboard, graphics system
and monitor but no local disk system. The system relies on disk resources
which are found in the network either on a dedicated server or shared over the
entire network resources.
divide-by-zero exception
The condition recognized by a processor that results from running a
program that attempts to divide by zero.
DNS
Domain Name Server is a hierarchical name service which maps high level
machine names to IP addresses.
double precision
Synonym for long-precision.
DWM
Diskless Workstation Manager is operating-system software that initializes
and maintains resources for diskless clients and diskless servers.
E
Ethernet
Ethernet is the standard hardware for TCP/IP LANs in the UNIX
marketplace. It is a 10 megabit per second baseband type network that uses the
contention based CSMA/CD (collision detect) media access method.
expression
A notation that represents a value: a primary appearing alone, or
combinations of primaries and operators. An expression can be arithmetic,
character, logical, or relational.
extent of a dimension
The number of different integer values that may be represented by
subscript expressions for a particular dimension in a subscript for an array.
external function
A function defined outside the program unit that refers to it. It may be
referred to in a procedure subprogram or in the main program, but it must not
refer to itself, either directly or indirectly. Contrast with statement
function.
EXTRINSIC (HPF)
Indicates that the procedure is HPF-conforming. Such a procedure is
referred to as an HPF procedure. This is the default; any procedure that does
not specify the EXTRINSIC attribute and is compiled with the -qhpf
option is considered to be an HPF procedure.
EXTRINSIC (HPF_LOCAL)
Indicates a procedure in HPF that is targeted to a single processor, with
many copies executing on different processors. Such a procedure is referred to
as a local procedure. The programming style in which many copies of the same
program run on multiple processors is often referred to as SPMD (single
program, multiple data.)
EXTRINSIC (HPF_SERIAL)
Indicates a procedure in HPF that is targeted to a single processor, with
only one instance of the procedure executing on only one processor. Such a
procedure is referred to as a serial procedure. (Serial procedures are useful
for code written in other languages or current XL Fortran programs that you do
not wish to recode.)
F
file
A set of related records treated as a unit, for example, in stock control,
a file could consist of a set of invoices.
file server
A centrally located computer that acts as a storehouse of data and
applications for numerous users of a local area network.
foreign host
Any host on the network other than the local host.
FTP
File transfer protocol.
function
In Fortran, a procedure that is invoked by referring to it in an
expression and that supplies a value to the expression. The value supplied is
the value of the function. See also external function, intrinsic function, and
statement function. Contrast with subroutine.
function reference
A Fortran source program reference to an intrinsic function, to an
external function, or to a statement function.
G
general matrix
A matrix with no assumed special properties such as symmetry. Synonym for
matrix.
global
(1) Pertaining to that which is defined in one subdivision of a computer
program and used in at least one other subdivision of the computer program.
(2) Pertaining to information available to more than one program or
subroutine. (3) Contrast with local.
H
home directory
The directory associated with an individual user.
host
A computer connected to a network, and providing an access method to that
network. A host provides end-user services.
I
integer constant
A string of decimal digits containing no decimal point and expressing a
whole number.
integer expression
An arithmetic expression whose values are of integer type.
integer type
An arithmetic data type capable of expressing the value of an integer. It
can have a positive, negative, or 0 value. It must not include a decimal
point.
Internet
The collection of worldwide networks and gateways which function as a
single, cooperative virtual network.
internet address
A unique 32-bit address assigned to hosts connected to a TCP/IP
network.
intrinsic function
A function, supplied by Fortran, that performs mathematical or character
operations.
IP
Internet protocol.
K
kernel
The core portion of the UNIX operating system which controls the resources
of the CPU and allocates them to the users. The kernel is memory-resident, is
said to run in "kernel mode" and is protected from user tampering by
the hardware.
L
LAN
Acronym for Local Area Network, a data network located on the user's
premises in which serial transmission is used for direct data communication
among data stations.
latency
The time interval between the instant at which an instruction control unit
initiates a call for data transmission and the instant at which the actual
transfer of data begins. Latency is related to the hardware characteristics of
the system and to the different layers of software that are involved in
initiating the task of packing and transmitting the data.
leading dimension
For a two-dimensional array, an increment used to find the starting point
for the matrix elements in each successive column of the array.
local
Pertaining to that which is defined and used only in one subdivision of a
computer program. Contrast with global.
local host
The computer to which a user's terminal is directly connected.
logical constant
A constant that can have one of two values: true or false. The form
of these values in Fortran is: .TRUE. and .FALSE. respectively.
logical expression
A logical primary alone or a combination of logical primaries and logical
operators. A logical expression can have one of two values: true or
false.
logical type
The data type for data items that can have the value true or false and
upon which logical operations such as .NOT. and .OR. can be performed. See
also "data type".
long-precision
Real type of data of length 8. Contrast with single precision and
short-precision.
M
main program
In Fortran, a program unit, required for running, that can call other
program units but cannot be called by them.
mainframe
A large computer to which other computers can be connected, so that they
can share facilities that the large computer provides; for example, it could
be a System/370 or System/390 computing system to which
personal computers are attached, so that they can upload and download programs
and data.
mask
To use a pattern of characters to control the retention or elimination of
portions of another pattern of characters. (I)
matrix
A rectangular array of elements, arranged in rows and columns, that may be
manipulated according to the rules of matrix algebra. (A) (I)
menu
A display of a list of available functions for selection by the user.
message passing
The method of communication among processor nodes operating in parallel
with distributed memory.
MPI
A Message Passing Interface standard.
MPMD (Multiple Program - Multiple Data)
A parallel programming model in which different, but related, programs are
run on different sets of data.
N
name
In Fortran, a string of up to six alphanumeric characters, the first of
which must be alphabetic. Used to identify a constant, a variable, an array, a
function, a subroutine, or a common block.
network
An interconnected group of nodes, lines, and terminals. A network provides
the ability to transmit data to and receive data from other systems and users.
NFS
Network file system. NFS allows different systems (UNIX or non-UNIX),
different architectures, or vendors connected to the same network, to access
remote files in a LAN environment as though they were local files.
node
In a network, the point where one or more functional units interconnect
transmission lines. A computer location defined in a network.
nodeid
The specific symbolic name assigned to a node during network definition.
O
overflow exception
A condition caused by the result of an arithmetic operation having a
magnitude that exceeds the largest possible number.
P
parallel processing
A multiprocessor architecture which allows processes to be allocated to
tightly coupled multiple processors in a cooperative processing environment,
allowing concurrent execution of tasks.
parameter
* (1) A variable that is given a constant value for a specified
application and that may denote the application. (2) An item in a menu for
which the operator specifies a value or for which the system provides a value
when the menu is interpreted. (3) A name in a procedure that is used to refer
to an argument that is passed to the procedure. (4) A particular piece of
information that a system or application program needs to process a request.
pipe
A UNIX utility allowing the output of one command to be the input of
another. Represented by the | symbol. It is also referred to as filtering
output.
port
(1) An endpoint for communication between devices, generally referring to
physical connection. (2) A 16-bit number identifying a particular TCP or UDP
resource within a given TCP/IP node.
primary
An irreducible unit of data; a single constant, variable, array element,
function reference, or expression enclosed in parentheses.
process
* (1) A unique, finite course of events defined by it purpose or by its
effect, achieved under defined conditions. * (2) Any operation or combination
of operations on data. * (3) A function being performed or waiting to be
performed. * (4) A program in operation. For example, a daemon is a system
process that is always running on the system.
process grid
A way to view a parallel machine as a logical one- or two-dimensional
rectangular grid of processes.
program exception
The condition recognized by a processor that results from running a
program that improperly specifies or uses instructions, operands, or control
information.
protocol
A set of semantic and syntactic rules that defines the behavior of
functional units in achieving communication.
PTF
Program Temporary Fix. A temporary solution or by-pass of a problem
diagnosed by IBM as resulting from a defect in a current unaltered release of
the program. A report of a problem caused by a suspected defect in a current
unaltered release of a program.
R
rank
In Fortran, the number of dimensions of an array. It is zero for scalar.
real constant
A string of decimal digits that expresses a real number. A real constant
must contain either a decimal point or a decimal exponent and may contain
both. For example, the real constant 0.36819E+2 has the value +36.819.
real type
An arithmetic data type, capable of approximating the value of a real
number. It can have a positive, negative, or 0 value.
remote host
See foreign host.
row-major order
A sequencing method used for storing multidimensional arrays according to
the subscripts of the array elements. In this method the rightmost subscript
position varies most rapidly and completes a full cycle before the next
subscript position to the left is incremented.
RISC
Reduced Instruction Set Computing (RISC), the technology for today's
high performance personal computers and workstations, was invented in 1975.
S
scalar
(1) A quantity characterized by a single number. (A) (I) (2) Contrast with
vector.
scope
(1) The portion of a computer program within which the definition of a
variable remains unchanged.
(2) In Appendix A. "BLACS Quick Reference Guide", for the broadcast topologies and global operations, scope can equal
'all', 'row', or 'column'.
server
(1) A function that provides services for users. A machine may run client
and server processes at the same time. (2) A machine that provides resources
to the network. It provides a network service, such as disk storage and file
transfer, or a program that uses such a service.
ScaLAPACK (Scalable Linear Algebra Package)
A scalable linear algebra library for distributed memory concurrent
computers. The library was jointly developed by the University of Tennessee,
Knoxville, Oak Ridge National Laboratory, and the University of California,
Berkeley.
shape of an array
The extents of all the dimensions of an array listed in order. For
example, the shape of a three-dimensional array that has four rows, five
columns, and three planes is (4,5,3) or 4 by 5 by 3.
shell
The shell is the primary user interface for the UNIX operating system. It
serves as command language interpreter, programming language, and allows
foreground and background processing. There are three different
implementations of the shell concept: Bourne, C and Korn.
short-precision
Real type data of length 4. Contrast with double precision and
long-precision.
single precision
Synonym for short-precision.
size of an array
The number of elements in an array. This is the product of the extents of
its dimensions.
SMIT
The System Management Interface Toolkit is a set of menu driven utilities
for AIX that provides functions such as transaction login, shell script
creation, automatic updates of object data base, etc.
SMP
Symmetric Multi-Processing.
SPMD (Single Program - Multiple Data)
A parallel programming model in which different processors execute the
same program on different sets of data.
statement
The basic unit of a program, that specifies an action to be performed, or
the nature and characteristics of the data to be processed, or information
about the program itself. Statements fall into two broad classes:
executable and nonexecutable.
statement function
A procedure specified by a single statement that is similar in form to an
arithmetic, logical, or character assignment statement. The statement must
appear after the specification statements and before the first executable
statement. In the remainder of the program it can be referenced as a function.
A statement function may be referred to only in the program unit in which it
is defined. Contrast with external function.
statement label
A number of from one through five decimal digits that is used to identify
a statement. Statement labels can be used to transfer control, to define the
range of a DO, or to refer to a FORMAT statement.
statement number
See "statement label".
stride
The increment used to step through array storage to select the vector or
matrix elements from the array.
subprogram
A program unit that is invoked by another program unit in the same
program. In Fortran, a subprogram has a FUNCTION, SUBROUTINE, or BLOCK DATA
statement as its first statement.
subscript
(1) A symbol that is associated with the name of a set to identify a
particular subset or element. (A) (2) A subscript expression or set of
subscript expressions, enclosed in parentheses and used with an array name to
identify a particular array element.
subscript expression
An integer expression in a subscript whose value and position in the
subscript determine the index number for the corresponding dimension in the
referenced array.
System Administrator
The user who is responsible for setting up, modifying, and maintaining the
computing system.
T
tar
Tape ARchive, is a standard UNIX data archive utility for storing data on
tape media.
TCP
Acronym for Transmission Control Protocol, a stream communication protocol
that includes error recovery and flow control.
TCP/IP
Acronym for Transmission Control Protocol/Internet Protocol, a
suite of protocols designed to allow communication between networks regardless
of the technologies implemented in each network.
Telnet
Terminal Emulation Protocol, a TCP/IP application protocol that
allows interactive access to foreign hosts.
thread
A thread is the element that is scheduled, and to which resources such as
execution time, locks, and queues may be assigned. There may be one or more
threads in a process, and each thread is executed by the operating system
concurrently.
thread-safe
A subroutine which may be called from multiple threads of the same process
simultaneously.
thread-tolerant
A library is thread-tolerant if it can be called from a single thread of a
multithreaded application. However, multiple simultaneous calls to the
thread-tolerant library from different threads of a single process causes
unpredictable results.
transaction
An exchange between the user and the system. Each activity the system
performs for the user is considered a transaction.
transfer
To send data from one place and to receive the data at another place.
Synonymous with move.
transmission
* The sending of data from one place for reception elsewhere.
type declaration
The explicit specification of the type of a constant, variable, array, or
function by use of an explicit type specification statement.
U
UDP
User Datagram Protocol.
underflow exception
A condition caused by the result of an arithmetic operation having a
magnitude less than the smallest possible nonzero number.
URL
Uniform Resource Locator.
user
Anyone who requires the services of a computing system.
V
variable
(1) A quantity that can assume any of a given set of values. (A) (2) A
data item, identified by a name, that is not a named constant, array, or array
element, and that can assume different values at different times during
program processing.
vector
A one-dimensional ordered collection of numbers.
W
working directory
A collection of files to be manipulated by an FTP operation.
workstation
A workstation is a single-user, high-performance microcomputer (or even a
minicomputer) which has been specialized in some way, usually for graphics
output. Such a machine has a screen and a keyboard, but is also capable of
extensive processing of your input before it is passed to the host. Likewise,
the host's responses may be extensively processed before being passed
along to your screen. A workstation may be intelligent enough to do much or
all the processing itself.
X
X Window System
A product developed at MIT that gives users windows into applications and
processes not located only or specifically on their own console or computer
system.