Unit-II

Parallel Computing
Introduction

Traditionally, software has been written for serial computation:

o To be run on a single computer having a single Central Processing Unit (CPU);
o A problem is broken into a discrete series of instructions.
o Instructions are executed one after another.
o Only one instruction may execute at any moment in time.

For example:

In the simplest sense, parallel computing is the simultaneous use of multiple compute
resources to solve a computational problem:
o To be run using multiple CPUs
o A problem is broken into discrete parts that can be solved concurrently

o
o

Each part is further broken down to a series of instructions

Instructions from each part execute simultaneously on different CPUs

For example:

The compute resources might be:

A single computer with multiple processors;

o An arbitrary number of computers connected by a network;
o A combination of both.
The computational problem should be able to:
o Be broken apart into discrete pieces of work that can be solved simultaneously;
o Execute multiple program instructions at any moment in time;
o Be solved in less time with multiple compute resources than with a single compute
resource.
o

Uses for Parallel Computing:

Science and Engineering: Historically, parallel computing has been considered to be

"the high end of computing", and has been used to model difficult problems in many
areas of science and engineering:
o
o

o
o

Atmosphere, Earth, Environment

Physics - applied, nuclear, particle,
condensed matter, high pressure, fusion,
photonics
Bioscience, Biotechnology, Genetics
Chemistry, Molecular Sciences

o
o
o
o

Geology, Seismology
Mechanical Engineering - from
prosthetics to spacecraft
Electrical Engineering, Circuit
Design, Microelectronics
Computer Science,
Mathematics

Industrial and Commercial: Today, commercial applications provide an equal or

greater driving force in the development of faster computers. These applications require
the processing of large amounts of data in sophisticated ways. For example:
o
o
o
o
o

Databases, data mining

Oil exploration
Web search engines, web based
business services
Medical imaging and diagnosis
Pharmaceutical design

o
o
o
o
o

Financial and economic modeling

Management of national and multi-national
corporations
Advanced graphics and virtual reality,
particularly in the entertainment industry
Networked video and multi-media
technologies
Collaborative work environments

Why Use Parallel Computing?

Main Reasons:
Save time and/or money: In theory, throwing more resources at a task will shorten its
time to completion, with potential cost savings. Parallel computers can be built from

cheap, commodity components.

Solve larger problems: Many problems are so large and/or complex that it is impractical
or impossible to solve them on a single computer, especially given limited computer
memory. For example:
o "Grand Challenge" problems requiring PetaFLOPS and PetaBytes of computing
resources.
o Web search engines/databases processing millions of transactions per second

Provide concurrency: A single compute resource can only do one thing at a time.
Multiple computing resources can be doing many things simultaneously. For example,
the Access Grid provides a global collaboration network where people from around the
world can meet and conduct work "virtually".

Use of non-local resources: Using compute resources on a wide area network, or even
the Internet when local compute resources are scarce. For example:
o SETI@home over 1.3 million users, 3.2 million computers in nearly every country
in the world.
o Folding@home uses over 450,000 cpus globally

Limits to serial computing: Both physical and practical reasons pose significant
constraints to simply building ever faster serial computers:
o Transmission speeds - the speed of a serial computer is directly dependent upon
how fast data can move through hardware. Absolute limits are the speed of light
(30 cm/nanosecond) and the transmission limit of copper wire (9 cm/nanosecond).
Increasing speeds necessitate increasing proximity of processing elements.
o Limits to miniaturization - processor technology is allowing an increasing number
of transistors to be placed on a chip. However, even with molecular or atomic-level
components, a limit will be reached on how small components can be.
o Economic limitations - it is increasingly expensive to make a single processor
faster. Using a larger number of moderately fast commodity processors to achieve
the same (or better) performance is less expensive.
o Current computer architectures are increasingly relying upon hardware level
parallelism to improve performance:
Multiple execution units
Pipelined instructions
Multi-core
Concepts related to parallel Computing

von Neumann Architecture

Named after the Hungarian mathematician John von Neumann who first authored the
general requirements for an electronic computer in his 1945 papers.
Since then, virtually all computers have followed this basic design, differing from earlier
computers which were programmed through "hard wiring".
o

o
o

Comprised of four main components:

Memory
Control Unit
Arithmetic Logic Unit
Input/Output
Read/write, random access memory is used
to store both program instructions and data
Program instructions are coded data
which tell the computer to do
something
Data is simply information to be
used by the program
Control unit fetches instructions/data from
memory, decodes the instructions and then
sequentially coordinates operations to
accomplish the programmed task.
Aritmetic Unit performs basic arithmetic
operations
Input/Output is the interface to the human
operator

Well, parallel computers still follow this basic design, just multiplied in units. The basic,
fundamental architecture remains the same.

Flynn's Classical Taxonomy

There are different ways to classify parallel computers. For example:

One of the more widely used classifications, in use since 1966, is called Flynn's
Taxonomy.
Flynn's taxonomy distinguishes multi-processor computer architectures according to
how they can be classified along the two independent dimensions of Instruction and
Data. Each of these dimensions can have only one of two possible states: Single or
Multiple.
The matrix below defines the 4 possible classifications according to Flynn:

SISD

SIMD

Single Instruction, Single Data

Single Instruction, Multiple Data

MISD

MIMD

Multiple Instruction, Single Data Multiple Instruction, Multiple Data

Single Instruction, Single Data (SISD):

A serial (non-parallel) computer

Single Instruction: Only one instruction stream is being acted
on by the CPU during any one clock cycle
Single Data: Only one data stream is being used as input during
any one clock cycle
Deterministic execution
This is the oldest and even today, the most common type of
computer
Examples: older generation mainframes, minicomputers and
workstations; most modern day PCs.

Single Instruction, Multiple Data (SIMD):

A type of parallel computer

Single Instruction: All processing units execute the same instruction at any given clock
cycle
Multiple Data: Each processing unit can operate on a different data element
Best suited for specialized problems characterized by a high degree of regularity, such as
graphics/image processing.
Synchronous (lockstep) and deterministic execution
Two varieties: Processor Arrays and Vector Pipelines
Examples:
o Processor Arrays: Connection Machine CM-2, MasPar MP-1 & MP-2, ILLIAC IV
o Vector Pipelines: IBM 9000, Cray X-MP, Y-MP & C90, Fujitsu VP, NEC SX-2, Hitachi
S820, ETA10
Most modern computers, particularly those with graphics processor units (GPUs)
employ SIMD instructions and execution units.

Multiple Instruction, Single Data (MISD):

A type of parallel computer

Multiple Instruction: Each processing unit operates on the data independently via
separate instruction streams.
Single Data: A single data stream is fed into multiple processing units.
Few actual examples of this class of parallel computer have ever existed. One is the
experimental Carnegie-Mellon C.mmp computer (1971).
Some conceivable uses might be:
o multiple frequency filters operating on a single signal stream
o multiple cryptography algorithms attempting to crack a single coded message.

Multiple Instruction, Multiple Data (MIMD):

A type of parallel computer

Multiple Instruction: Every processor may be executing a different instruction stream
Multiple Data: Every processor may be working with a different data stream
Execution can be synchronous or asynchronous, deterministic or non-deterministic
Currently, the most common type of parallel computer - most modern supercomputers
fall into this category.
Examples: most current supercomputers, networked parallel computer clusters and
"grids", multi-processor SMP computers, multi-core PCs.

Note: many MIMD architectures also include SIMD execution sub-components

Parallel Terminology
Like everything else, parallel computing has its own "jargon". Some of the more commonly
used terms associated with parallel computing are listed below. Most of these will be
discussed in more detail later.
Supercomputing / High Performance Computing (HPC)
Using the world's fastest and largest computers to solve large problems.
Node
A standalone "computer in a box". Usually comprised of multiple
CPUs/processors/cores. Nodes are networked together to comprise a supercomputer.
CPU / Socket / Processor / Core
This varies, depending upon who you talk to. In the past, a CPU (Central Processing Unit)
was a singular execution component for a computer. Then, multiple CPUs were
incorporated into a node. Then, individual CPUs were subdivided into multiple "cores",
each being a unique execution unit. CPUs with multiple cores are sometimes called
"sockets" - vendor dependent. The result is a node with multiple CPUs, each containing
multiple cores. The nomenclature is confused at times. Wonder why?

Task
A logically discrete section of computational work. A task is typically a program or
program-like set of instructions that is executed by a processor. A parallel program
consists of multiple tasks running on multiple processors.
Pipelining
Breaking a task into steps performed by different processor units, with inputs streaming
through, much like an assembly line; a type of parallel computing.
Shared Memory
From a strictly hardware point of view, describes a computer architecture where all
processors have direct (usually bus based) access to common physical memory. In a
programming sense, it describes a model where parallel tasks all have the same
"picture" of memory and can directly address and access the same logical memory
locations regardless of where the physical memory actually exists.
Symmetric Multi-Processor (SMP)
Hardware architecture where multiple processors share a single address space and
access to all resources; shared memory computing.

Distributed Memory
In hardware, refers to network based memory access for physical memory that is not
common. As a programming model, tasks can only logically "see" local machine memory
and must use communications to access memory on other machines where other tasks
are executing.
Communications
Parallel tasks typically need to exchange data. There are several ways this can be
accomplished, such as through a shared memory bus or over a network, however the
actual event of data exchange is commonly referred to as communications regardless of
the method employed.
Synchronization
The coordination of parallel tasks in r eal time, very often associated with
communications. Often implemented by establishing a synchronization point within an
application where a task may not proceed further until another task(s) reaches the same
or logically equivalent point.
Synchronization usually involves waiting by at least one task, and can therefore cause a
parallel application's wall clock execution time to increase.
Granularity
In parallel computing, granularity is a qualitative measure of the ratio of computation to
communication.

Coarse: relatively large amounts of computational work are done between

communication events
Fine: relatively small amounts of computational work are done between
communication events

Observed Speedup
Observed speedup of a code which has been parallelized, defined as:
wall-clock time of serial execution
----------------------------------wall-clock time of parallel execution

One of the simplest and most widely used indicators for a parallel program's
performance.
Parallel Overhead
The amount of time required to coordinate parallel tasks, as opposed to doing useful
work. Parallel overhead can include factors such as:

Task start-up time

Synchronizations
Data communications
Software overhead imposed by parallel compilers, libraries, tools, operating
system, etc.
Task termination time

Massively Parallel
Refers to the hardware that comprises a given parallel system - having many processors.
The meaning of "many" keeps increasing, but currently, the largest parallel computers
can be comprised of processors numbering in the hundreds of thousands.
Embarrassingly Parallel
Solving many similar, but independent tasks simultaneously; little to no need for
coordination between the tasks.
Scalability
Refers to a parallel system's (hardware and/or software) ability to demonstrate a
proportionate increase in parallel speedup with the addition of more processors.
Factors that contribute to scalability include:

Hardware - particularly memory-cpu bandwidths and network communications

Application algorithm
Parallel overhead related
Characteristics of your specific application and coding

Parallel Computer Memory Architectures

Shared Memory

General Characteristics:

Shared memory parallel computers vary widely, but generally have in common the
ability for all processors to access all memory as global address space.
Multiple processors can operate independently but share the same memory resources.
Changes in a memory location effected by one processor are visible to all other
processors.
Shared memory machines can be divided into two main classes based upon memory
access times: UMA and NUMA.

Uniform Memory Access (UMA):

Most commonly represented today by Symmetric Multiprocessor (SMP) machines

Identical processors
Equal access and access times to memory
Sometimes called CC-UMA - Cache Coherent UMA. Cache coherent means if one
processor updates a location in shared memory, all the other processors know about
the update. Cache coherency is accomplished at the hardware level.

Non-Uniform Memory Access (NUMA):

Often made by physically linking two or more SMPs

One SMP can directly access memory of another SMP
Not all processors have equal access time to all memories
Memory access across link is slower
If cache coherency is maintained, then may also be called CC-NUMA - Cache Coherent
NUMA

Advantages:

Global address space provides a user-friendly programming perspective to memory

Data sharing between tasks is both fast and uniform due to the proximity of memory to
CPUs

Disadvantages:

Primary disadvantage is the lack of scalability between memory and CPUs. Adding more
CPUs can geometrically increases traffic on the shared memory-CPU path, and for
cache coherent systems, geometrically increase traffic associated with cache/memory
management.
Programmer responsibility for synchronization constructs that ensure "correct" access
of global memory.
Expense: it becomes increasingly difficult and expensive to design and produce shared
memory machines with ever increasing numbers of processors.
Distributed Memory

General Characteristics:

Like shared memory systems, distributed memory systems vary widely but share a
common characteristic. Distributed memory systems require a communication network
to connect inter-processor memory.

Processors have their own local memory. Memory addresses in one processor do not
map to another processor, so there is no concept of global address space across all
processors.
Because each processor has its own local memory, it operates independently. Changes
it makes to its local memory have no effect on the memory of other processors. Hence,
the concept of cache coherency does not apply.
When a processor needs access to data in another processor, it is usually the task of
the programmer to explicitly define how and when data is communicated.
Synchronization between tasks is likewise the programmer's responsibility.
The network "fabric" used for data transfer varies widely, though it can can be as
simple as Ethernet.

Advantages:

Memory is scalable with the number of processors. Increase the number of processors
and the size of memory increases proportionately.
Each processor can rapidly access its own memory without interference and without
the overhead incurred with trying to maintain cache coherency.
Cost effectiveness: can use commodity, off-the-shelf processors and networking.

Disadvantages:

The programmer is responsible for many of the details associated with data
communication between processors.
It may be difficult to map existing data structures, based on global memory, to this
memory organization.
Non-uniform memory access (NUMA) times

Hybrid Distributed-Shared Memory

The largest and fastest computers in the world today employ both shared and
distributed memory architectures.

The shared memory component can be a cache coherent SMP machine and/or graphics
processing units (GPU).
The distributed memory component is the networking of multiple SMP/GPU machines,
which know only about their own memory - not the memory on another machine.
Therefore, network communications are required to move data from one SMP/GPU to
another.
Current trends seem to indicate that this type of memory architecture will continue to
prevail and increase at the high end of computing for the foreseeable future.
Advantages and Disadvantages: whatever is common to both shared and distributed
memory architectures.

Parallel Programming Models

Overview

There are several parallel programming models in common use:

o Shared Memory (without threads)
o Threads
o Distributed Memory / Message Passing
o Data Parallel
o Hybrid
o Single Program Multiple Data (SPMD)
o Multiple Program Multiple Data (MPMD)
Parallel programming models exist as an abstraction above hardware and memory
architectures.

Which model to use? This is often a combination of what is available and personal
choice. There is no "best" model, although there certainly are better implementations
of some models over others.
The following sections describe each of the models mentioned above, and also discuss
some of their actual implementations.

Shared Memory Model (without threads)

In this programming model, tasks share a common address space, which they read and
write to asynchronously.
Various mechanisms such as locks / semaphores may be used to control access to the
shared memory.
An advantage of this model from the programmer's point of view is that the notion of
data "ownership" is lacking, so there is no need to specify explicitly the communication
of data between tasks. Program development can often be simplified.
An important disadvantage in terms of performance is that it becomes more difficult to
understand and manage data locality.
o Keeping data local to the processor that works on it conserves memory accesses,
cache refreshes and bus traffic that occurs when multiple processors use the
same data.
o Unfortunately, controlling data locality is hard to understand and beyond the
control of the average user.

Implementations:

Native compilers and/or hardware translate user program variables into actual memory
addresses, which are global. On stand-alone SMP machines, this is straightforward.
On distributed shared memory machines, such as the SGI Origin, memory is physically
distributed across a network of machines, but made global through specialized
hardware and software.

Threads Model

This programming model is a type of shared memory programming.

In the threads model of parallel programming, a single process can have multiple,
concurrent execution paths.

Perhaps the most simple analogy that

can be used to describe threads is the
concept of a single program that
includes a number of subroutines:
o The main program a.out is
scheduled to run by the native
operating system. a.out loads
and acquires all of the necessary
system and user resources to run.
o a.out performs some serial
work, and then creates a number
of tasks (threads) that can be
scheduled and run by the operating system concurrently.
o Each thread has local data, but also, shares the entire resources of a.out. This
saves the overhead associated with replicating a program's resources for each
thread. Each thread also benefits from a global memory view because it shares
the memory space of a.out.
o A thread's work may best be described as a subroutine within the main program.
Any thread can execute any subroutine at the same time as other threads.
o Threads communicate with each other through global memory (updating address
locations). This requires synchronization constructs to ensure that more than one
thread is not updating the same global address at any time.
o Threads can come and go, but a.out remains present to provide the necessary
shared resources until the application has completed.

Implementations:

From a programming perspective, threads implementations commonly comprise:

o A library of subroutines that are called from within parallel source code
o A set of compiler directives imbedded in either serial or parallel source code
In both cases, the programmer is responsible for determining all parallelism.

Threaded implementations are not new in computing. Historically, hardware vendors

have implemented their own proprietary versions of threads. These implementations
differed substantially from each other making it difficult for programmers to develop
portable threaded applications.
Unrelated standardization efforts have resulted in two very different implementations
of threads: POSIX Threads and OpenMP.
POSIX Threads
o Library based; requires parallel coding

Specified by the IEEE POSIX 1003.1c standard (1995).

o C Language only
o Commonly referred to as Pthreads.
o Most hardware vendors now offer Pthreads in addition to their proprietary
threads implementations.
o Very explicit parallelism; requires significant programmer attention to detail.
OpenMP
o Compiler directive based; can use serial code
o Jointly defined and endorsed by a group of major computer hardware and
software vendors. The OpenMP Fortran API was released October 28, 1997. The
C/C++ API was released in late 1998.
o Portable / multi-platform, including Unix and Windows NT platforms
o Available in C/C++ and Fortran implementations
o Can be very easy and simple to use - provides for "incremental parallelism"
Microsoft has its own implementation for threads, which is not related to the UNIX
POSIX standard or OpenMP.
o

Distributed Memory /
Message Passing Model

This model demonstrates

the following
characteristics:
o A set of tasks that
use their own local
memory during
computation.
Multiple tasks can
reside on the same
physical machine
and/or across an
arbitrary number of
machines.
o Tasks exchange data
through communications by sending and receiving messages.
o Data transfer usually requires cooperative operations to be performed by each
process. For example, a send operation must have a matching receive operation.

Implementations:

From a programming perspective, message passing implementations usually comprise a

library of subroutines. Calls to these subroutines are imbedded in source code. The
programmer is responsible for determining all parallelism.
Historically, a variety of message passing libraries have been available since the 1980s.
These implementations differed substantially from each other making it difficult for
programmers to develop portable applications.
In 1992, the MPI Forum was formed with the primary goal of establishing a standard
interface for message passing implementations.
Part 1 of the Message Passing Interface (MPI) was released in 1994. Part 2 (MPI-2) was
released in 1996.
MPI is now the "de facto" industry standard for message passing, replacing virtually all
other message passing implementations used for production work. MPI
implementations exist for virtually all popular parallel computing platforms. Not all
implementations include everything in both MPI1 and MPI2.

Data Parallel Model

The data parallel model

demonstrates the following
characteristics:
o Most of the parallel work
focuses on performing
operations on a data set.
The data set is typically
organized into a common
structure, such as an
array or cube.
o A set of tasks work
collectively on the same
data structure, however,
each task works on a
different partition of the
same data structure.
o Tasks perform the same
operation on their partition of work, for example, "add 4 to every array
element".
On shared memory architectures, all tasks may have access to the data structure
through global memory. On distributed memory architectures the data structure is split
up and resides as "chunks" in the local memory of each task.

Implementations:

Programming with the data parallel model is usually accomplished by writing a program
with data parallel constructs. The constructs can be calls to a data parallel subroutine
library or, compiler directives recognized by a data parallel compiler.
Fortran 90 and 95 (F90, F95): ISO/ANSI standard extensions to Fortran 77.
o Contains everything that is in Fortran 77
o New source code format; additions to character set
o Additions to program structure and commands
o Variable additions - methods and arguments
o Pointers and dynamic memory allocation added
o Array processing (arrays treated as objects) added
o Recursive and new intrinsic functions added
o Many other new features
Implementations are available for most common parallel platforms.

High Performance Fortran (HPF): Extensions to Fortran 90 to support data parallel

programming.
o Contains everything in Fortran 90
o Directives to tell compiler how to distribute data added
o Assertions that can improve optimization of generated code added
o Data parallel constructs added (now part of Fortran 95)
HPF compilers were relatively common in the 1990s, but are no longer commonly
implemented.

Compiler Directives: Allow the programmer to specify the distribution and alignment of
data. Fortran implementations are available for most common parallel platforms.
Distributed memory implementations of this model usually require the compiler to
produce object code with calls to a message passing library (MPI) for data distribution.
All message passing is done invisibly to the programmer.

Hybrid Model

A hybrid model
combines more than
one of the previously
described
programming models.
Currently, a common
example of a hybrid
model is the
combination of the
message passing
model (MPI) with the
threads model
(OpenMP).
o Threads perform computationally intensive kernels using local, on-node data
o Communications between processes on different nodes occurs over the network
using MPI
This hybrid model lends itself well to the increasingly common hardware environment
of clustered multi/many-core machines.
Another similar and increasingly popular example of a hybrid model is using MPI with
GPU (Graphics Processing Unit) programming.
o GPUs perform computationally intensive kernels using local, on-node data
o Communications between processes on different nodes occurs over the network
using MPI

SPMD and MPMD

Single Program Multiple Data (SPMD):

SPMD is actually a "high level"

programming model that can be
built upon any combination of
the previously mentioned parallel
programming models.
SINGLE PROGRAM: All tasks
execute their copy of the same
program simultaneously. This program can be threads, message passing, data parallel
or hybrid.
MULTIPLE DATA: All tasks may use different data
SPMD programs usually have the necessary logic programmed into them to allow
different tasks to branch or conditionally execute only those parts of the program they

are designed to execute. That is, tasks do not necessarily have to execute the entire
program - perhaps only a portion of it.
The SPMD model, using message passing or hybrid programming, is probably the most
commonly used parallel programming model for multi-node clusters.

Multiple Program Multiple Data (MPMD):

Like SPMD, MPMD is actually a

"high level" programming model
that can be built upon any
combination of the previously
mentioned parallel programming
models.
MULTIPLE PROGRAM: Tasks may
execute different programs simultaneously. The programs can be threads, message
passing, data parallel or hybrid.
MULTIPLE DATA: All tasks may use different data
MPMD applications are not as common as SPMD applications, but may be better suited
for certain types of problems, particularly those that lend themselves better to
functional decomposition than domain decomposition (discussed later under
Partioning).

Designing Parallel Programs

Automatic vs. Manual Parallelization

Designing and developing parallel programs has characteristically been a very manual
process. The programmer is typically responsible for both identifying and actually
implementing parallelism.
Very often, manually developing parallel codes is a time consuming, complex, errorprone and iterative process.
For a number of years now, various tools have been available to assist the programmer
with converting serial programs into parallel programs. The most common type of tool
used to automatically parallelize a serial program is a parallelizing compiler or preprocessor.
A parallelizing compiler generally works in two different ways:
o Fully Automatic
o The compiler analyzes the source code and identifies opportunities for
parallelism.
o The analysis includes identifying inhibitors to parallelism and possibly a
cost weighting on whether or not the parallelism would actually improve
performance.

Loops (do, for) loops are the most frequent target for automatic
parallelization.
o Programmer Directed
o Using "compiler directives" or possibly compiler flags, the programmer
explicitly tells the compiler how to parallelize the code.
o May be able to be used in conjunction with some degree of automatic
parallelization also.
If you are beginning with an existing serial code and have time or budget constraints,
then automatic parallelization may be the answer. However, there are several important
caveats that apply to automatic parallelization:
o Wrong results may be produced
o Performance may actually degrade
o Much less flexible than manual parallelization
o Limited to a subset (mostly loops) of code
o May actually not parallelize code if the analysis suggests there are inhibitors or
the code is too complex
The remainder of this section applies to the manual method of developing parallel codes.
o

Understand the Problem and the Program

Undoubtedly, the first step in developing parallel software is to first understand the
problem that you wish to solve in parallel. If you are starting with a serial program, this
necessitates understanding the existing code also.
Before spending time in an attempt to develop a parallel solution for a problem,
determine whether or not the problem is one that can actually be parallelized.
o Example of Parallelizable Problem:
Calculate the potential energy for each of several thousand
independent conformations of a molecule. When done, find the
minimum energy conformation.
o

This problem is able to be solved in parallel. Each of the molecular conformations

is independently determinable. The calculation of the minimum energy
conformation is also a parallelizable problem.
Example of a Non-parallelizable Problem:
Calculation of the Fibonacci series (0,1,1,2,3,5,8,13,21,...) by use
of the formula:
F(n) = F(n-1) + F(n-2)

This is a non-parallelizable problem because the calculation of the Fibonacci

sequence as shown would entail dependent calculations rather than independent

ones. The calculation of the F(n) value uses those of both F(n-1) and F(n-2).
These three terms cannot be calculated independently and therefore, not in
parallel.

Identify the program's hotspots:

o Know where most of the real work is being done. The majority of scientific and
technical programs usually accomplish most of their work in a few places.
o Profilers and performance analysis tools can help here
o Focus on parallelizing the hotspots and ignore those sections of the program that
account for little CPU usage.
Identify bottlenecks in the program
o Are there areas that are disproportionately slow, or cause parallelizable work to
halt or be deferred? For example, I/O is usually something that slows a program
down.
o May be possible to restructure the program or use a different algorithm to reduce
or eliminate unnecessary slow areas
Identify inhibitors to parallelism. One common class of inhibitor is data dependence, as
demonstrated by the Fibonacci sequence above.
Investigate other algorithms if possible. This may be the single most important
consideration when designing a parallel application.
Take advantage of optimized third party parallel software and highly optimized math
libraries available from leading vendors (IBM's ESSL, Intel's MKL, AMD's AMCL,
etc.).

Partitioning

One of the first steps in designing a parallel program is to break the problem into
discrete "chunks" of work that can be distributed to multiple tasks. This is known as
decomposition or partitioning.
There are two basic ways to partition computational work among parallel tasks: domain
decomposition and functional decomposition.

Domain Decomposition:

In this type of partitioning, the data associated with a problem is decomposed. Each
parallel task then works on a portion of of the data.

There are different ways to partition data: 0ne Dimensional and two Dimensional

Functional Decomposition:

In this approach, the focus is on the computation that is to be performed rather than on
the data manipulated by the computation. The problem is decomposed according to the
work that must be done. Each task then performs a portion of the overall work.

Functional decomposition lends itself well to problems that can be split into different
tasks. For example:
Ecosystem Modeling
Each program calculates the population of a given group, where each group's growth
depends on that of its neighbors. As time progresses, each process calculates its current
state, then exchanges information with the neighbor populations. All tasks then progress

to calculate the state at the next time step.

Signal Processing
An audio signal data set is passed through four distinct computational filters. Each filter
is a separate process. The first segment of data must pass through the first filter before
progressing to the second. When it does, the second segment of data passes through the
first filter. By the time the fourth segment of data is in the first filter, all four tasks are
busy.

Climate Modeling
Each model component can be thought of as a separate task. Arrows represent exchanges
of data between components during computation: the atmosphere model generates wind
velocity data that are used by the ocean model, the ocean model generates sea surface
temperature data that are used by the atmosphere model, and so on.

Combining these two types of problem decomposition is common and natural.

Communications
Who Needs Communications?
The need for communications between tasks depends upon your problem:

You DON'T need communications

o Some types of problems can be decomposed and executed in parallel with
virtually no need for tasks to share data. For example, imagine an image
processing operation where every pixel in a black and white image needs to have
its color reversed. The image data can easily be distributed to multiple tasks that
then act independently of each other to do their portion of the work.
o These types of problems are often called embarrassingly parallel because they are
so straight-forward. Very little inter-task communication is required.
You DO need communications
o Most parallel applications are not quite so simple, and do require tasks to share
data with each other. For example, a 3-D heat diffusion problem requires a task to
know the temperatures calculated by the tasks that have neighboring data.
Changes to neighboring data has a direct effect on that task's data.

Factors to Consider:
There are a number of important factors to consider when designing your program's
inter-task communications:

Cost of communications
o Inter-task communication virtually always implies overhead.
o Machine cycles and resources that could be used for computation are instead used

to package and transmit data.

o Communications frequently require some type of synchronization between tasks,
which can result in tasks spending time "waiting" instead of doing work.
o Competing communication traffic can saturate the available network bandwidth,
further aggravating performance problems.
Latency vs. Bandwidth
o latency is the time it takes to send a minimal (0 byte) message from point A to
point B. Commonly expressed as microseconds.
o bandwidth is the amount of data that can be communicated per unit of time.
Commonly expressed as megabytes/sec or gigabytes/sec.
o Sending many small messages can cause latency to dominate communication
overheads. Often it is more efficient to package small messages into a larger
message, thus increasing the effective communications bandwidth.
Visibility of communications
o With the Message Passing Model, communications are explicit and generally
quite visible and under the control of the programmer.
o With the Data Parallel Model, communications often occur transparently to the
programmer, particularly on distributed memory architectures. The programmer
may not even be able to know exactly how inter-task communications are being
accomplished.
Synchronous vs. asynchronous communications
o Synchronous communications require some type of "handshaking" between tasks
that are sharing data. This can be explicitly structured in code by the programmer,
or it may happen at a lower level unknown to the programmer.
o Synchronous communications are often referred to as blocking communications
since other work must wait until the communications have completed.
o Asynchronous communications allow tasks to transfer data independently from
one another. For example, task 1 can prepare and send a message to task 2, and
then immediately begin doing other work. When task 2 actually receives the data
doesn't matter.
o Asynchronous communications are often referred to as non-blocking
communications since other work can be done while the communications are
taking place.
o Interleaving computation with communication is the single greatest benefit for
using asynchronous communications.
Scope of communications
o Knowing which tasks must communicate with each other is critical during the
design stage of a parallel code. Both of the two scopings described below can be
implemented synchronously or asynchronously.
o Point-to-point - involves two tasks with one task acting as the sender/producer of
data, and the other acting as the receiver/consumer.
o Collective - involves data sharing between more than two tasks, which are often

specified as being members in a common group, or collective. Some common

variations (there are more):

Efficiency of communications
o Very often, the programmer will have a choice with regard to factors that can
affect communications performance. Only a few are mentioned here.
o Which implementation for a given model should be used? Using the Message
Passing Model as an example, one MPI implementation may be faster on a given
hardware platform than another.
o What type of communication operations should be used? As mentioned
previously, asynchronous communication operations can improve overall program
performance.
o Network media - some platforms may offer more than one network for
communications. Which one is best?
Overhead and Complexity
Finally, realize that this is only a partial list of things to consider!!!

Synchronization
Types of Synchronization:

Barrier
o Usually implies that all tasks are involved
o Each task performs its work until it reaches the barrier. It then stops, or "blocks".
o When the last task reaches the barrier, all tasks are synchronized.
o What happens from here varies. Often, a serial section of work must be done. In
other cases, the tasks are automatically released to continue their work.

Lock / semaphore
o Can involve any number of tasks
o Typically used to serialize (protect) access to global data or a section of code.
Only one task at a time may use (own) the lock / semaphore / flag.
o The first task to acquire the lock "sets" it. This task can then safely (serially)
access the protected data or code.
o Other tasks can attempt to acquire the lock but must wait until the task that owns
the lock releases it.
o Can be blocking or non-blocking
Synchronous communication operations
o Involves only those tasks executing a communication operation
o When a task performs a communication operation, some form of coordination is
required with the other task(s) participating in the communication. For example,
before a task can perform a send operation, it must first receive an
acknowledgment from the receiving task that it is OK to send.
o Discussed previously in the Communications section.

Data Dependencies
Definition:

A dependence exists between program statements when the order of statement execution
affects the results of the program.
A data dependence results from multiple use of the same location(s) in storage by
different tasks.
Dependencies are important to parallel programming because they are one of the
primary inhibitors to parallelism.

Examples:

Loop carried data dependence

DO 500 J = MYSTART,MYEND
A(J) = A(J-1) * 2.0
500 CONTINUE
The value of A(J-1) must be computed before the value of A(J), therefore A(J) exhibits a
data dependency on A(J-1). Parallelism is inhibited.
If Task 2 has A(J) and task 1 has A(J-1), computing the correct value of A(J)
necessitates:

Distributed memory architecture - task 2 must obtain the value of A(J-1) from task
1 after task 1 finishes its computation
o Shared memory architecture - task 2 must read A(J-1) after task 1 updates it
Loop independent data dependence
o

task 1
------

task 2
------

X = 2
.
.
Y = X**2

X = 4
.
.
Y = X**3

As with the previous example, parallelism is inhibited. The value of Y is dependent on:
Distributed memory architecture - if or when the value of X is communicated
between the tasks.
o Shared memory architecture - which task last stores the value of X.
Although all data dependencies are important to identify when designing parallel
programs, loop carried dependencies are particularly important since loops are possibly
the most common target of parallelization efforts.
o

How to Handle Data Dependencies:

Distributed memory architectures - communicate required data at synchronization

points.
Shared memory architectures -synchronize read/write operations between tasks.

Load Balancing

Load balancing refers to the practice of distributing work among tasks so that all tasks
are kept busy all of the time. It can be considered a minimization of task idle time.
Load balancing is important to parallel programs for performance reasons. For example,
if all tasks are subject to a barrier synchronization point, the slowest task will determine
the overall performance.

How to Achieve Load Balance:

Equally partition the work each task receives

o For array/matrix operations where each task performs similar work, evenly
distribute the data set among the tasks.
o For loop iterations where the work done in each iteration is similar, evenly
distribute the iterations across the tasks.
o If a heterogeneous mix of machines with varying performance characteristics are
being used, be sure to use some type of performance analysis tool to detect any
load imbalances. Adjust work accordingly.
Use dynamic work assignment
o Certain classes of problems result in load imbalances even if data is evenly
distributed among tasks:
o Sparse arrays - some tasks will have actual data to work on while others
have mostly "zeros".
o Adaptive grid methods - some tasks may need to refine their mesh while
others don't.
o N-body simulations - where some particles may migrate to/from their
original task domain to another task's; where the particles owned by some
tasks require more work than those owned by other tasks.
o When the amount of work each task will perform is intentionally variable, or is
unable to be predicted, it may be helpful to use a scheduler - task pool approach.
As each task finishes its work, it queues to get a new piece of work.
o It may become necessary to design an algorithm which detects and handles load
imbalances as they occur dynamically within the code.

Granularity
Computation / Communication Ratio:

In parallel computing, granularity is a qualitative measure of the ratio of computation to

communication.

Periods of computation are typically separated from periods of communication by

synchronization events.
Fine-grain Parallelism:

Relatively small amounts of computational work are done between communication

events
Low computation to communication ratio
Facilitates load balancing
Implies high communication overhead and less opportunity for performance
enhancement
If granularity is too fine it is possible that the overhead required for communications and
synchronization between tasks takes longer than the computation.

Coarse-grain Parallelism:

Relatively large amounts of computational work are done between

communication/synchronization events
High computation to communication ratio
Implies more opportunity for performance increase
Harder to load balance efficiently

Which is Best?

The most efficient granularity is dependent on the algorithm and the hardware
environment in which it runs.
In most cases the overhead associated with communications and synchronization is high
relative to execution speed so it is advantageous to have coarse granularity.
Fine-grain parallelism can help reduce overheads due to load imbalance.

I/O
The Bad News:

I/O operations are generally regarded as inhibitors to parallelism

Parallel I/O systems may be immature or not available for all platforms
In an environment where all tasks see the same file space, write operations can result in
file overwriting
Read operations can be affected by the file server's ability to handle multiple read
requests at the same time
I/O that must be conducted over the network (NFS, non-local) can cause severe
bottlenecks and even crash file servers.

The Good News:

Parallel file systems are available. For example:

o GPFS: General Parallel File System for AIX (IBM)
o Lustre: for Linux clusters (Oracle)
o PVFS/PVFS2: Parallel Virtual File System for Linux clusters
(Clemson/Argonne/Ohio State/others)
o PanFS: Panasas ActiveScale File System for Linux clusters (Panasas, Inc.)
o HP SFS: HP StorageWorks Scalable File Share. Lustre based parallel file system
(Global File System for Linux) product from HP
The parallel I/O programming interface specification for MPI has been available since
1996 as part of MPI-2. Vendor and "free" implementations are now commonly available.
A few pointers:
o Rule #1: Reduce overall I/O as much as possible
o If you have access to a parallel file system, investigate using it.
o Writing large chunks of data rather than small packets is usually significantly
more efficient.
o Confine I/O to specific serial portions of the job, and then use parallel
communications to distribute data to parallel tasks. For example, Task 1 could
read an input file and then communicate required data to other tasks. Likewise,
Task 1 could perform write operation after receiving required data from all other
tasks.
o Use local, on-node file space for I/O if possible. For example, each node may
have /tmp filespace which can used. This is usually much more efficient than
performing I/O over the network to one's home directory.

Limits and Costs of Parallel Programming

Amdahl's Law:

Amdahl's Law states that potential

program speedup is defined by the
fraction of code (P) that can be
parallelized:

speedup

1
-----

--1

If none of the code can be

parallelized, P = 0 and the speedup =
1 (no speedup).
If all of the code is parallelized, P = 1
and the speedup is infinite (in
theory).
If 50% of the code can be
parallelized, maximum speedup = 2,
meaning the code will run twice as
fast.
Introducing the number of processors
performing the parallel fraction of
work, the relationship can be
modeled by:

1
speedup
-------

----P

S
--N

where P = parallel fraction, N =

number of processors and S = serial

fraction.

It soon becomes obvious that there

are limits to the scalability of
parallelism. For example:

speedup
------------------------------N
P = .50
P = .90
P = .99
----------------------10
1.82
5.26
9.17
100
1.98
9.17
50.25
1000
1.99
9.91
90.99
10000
1.99
9.91
99.02
100000
1.99
9.99
99.90

However, certain problems demonstrate increased performance by increasing the

problem size. For example:

2D Grid Calculations
Serial fraction

85 seconds
15 seconds

85%
15%

We can increase the problem size by doubling the grid dimensions and halving the time
step. This results in four times the number of grid points and twice the number of time
steps. The timings then look like:
2D Grid Calculations
Serial fraction

680 seconds
15 seconds

97.84%
2.16%

Problems that increase the percentage of parallel time with their size are more scalable
than problems with a fixed percentage of parallel time.

Complexity:

In general, parallel applications are much more complex than corresponding serial
applications, perhaps an order of magnitude. Not only do you have multiple instruction
streams executing at the same time, but you also have data flowing between them.
The costs of complexity are measured in programmer time in virtually every aspect of
the software development cycle:
o Design
o Coding
o Debugging
o Tuning
o Maintenance
Adhering to "good" software development practices is essential when when working
with parallel applications - especially if somebody besides you will have to work with
the software.

Portability:

Thanks to standardization in several APIs, such as MPI, POSIX threads, HPF and
OpenMP, portability issues with parallel programs are not as serious as in years past.
However...
All of the usual portability issues associated with serial programs apply to parallel
programs. For example, if you use vendor "enhancements" to Fortran, C or C++,
portability will be a problem.
Even though standards exist for several APIs, implementations will differ in a number of
details, sometimes to the point of requiring code modifications in order to effect
portability.
Operating systems can play a key role in code portability issues.
Hardware architectures are characteristically highly variable and can affect portability.

Resource Requirements:

The primary intent of parallel programming is to decrease execution wall clock time,
however in order to accomplish this, more CPU time is required. For example, a parallel
code that runs in 1 hour on 8 processors actually uses 8 hours of CPU time.
The amount of memory required can be greater for parallel codes than serial codes, due
to the need to replicate data and for overheads associated with parallel support libraries
and subsystems.
For short running parallel programs, there can actually be a decrease in performance
compared to a similar serial implementation. The overhead costs associated with setting
up the parallel environment, task creation, communications and task termination can
comprise a significant portion of the total execution time for short runs.

Scalability:

The ability of a parallel program's performance to scale is a result of a number of

interrelated factors. Simply adding more machines is rarely the answer.
The algorithm may have inherent limits to scalability. At some point, adding more
resources causes performance to decrease. Most parallel solutions demonstrate this
characteristic at some point.
Hardware factors play a significant role in scalability. Examples:
o Memory-cpu bus bandwidth on an SMP machine
o Communications network bandwidth
o Amount of memory available on any given machine or set of machines
o Processor clock speed
Parallel support libraries and subsystems software can limit scalability independent of
your application.

Performance Analysis and Tuning

As with debugging, monitoring and analyzing parallel program execution is significantly

more of a challenge than for serial programs.
A number of parallel tools for execution monitoring and program analysis are available.
Some are quite useful; some are cross-platform also.
Some starting points:
o LC's "Supported Software and Computing Tools" web pages at:
computing.llnl.gov/code/content/software_tools.php
o A dated, but potentially useful LC whitepaper on the subject of "High
Performance Tools and Technologies" describes a large number of tools, and a
number of performance related topics applicable to code developers. Find it at:
computing.llnl.gov/tutorials/performance_tools/HighPerformanceToolsTechnolog
iesLC.pdf.
o Performance Analysis Tools Tutorial
Work remains to be done, particularly in the area of scalability.

Parallel Examples
Array Processing

This example demonstrates calculations on 2dimensional array elements, with the

computation on each array element being
independent from other array elements.
The serial program calculates one element at a
time in sequential order.
Serial code could be of the form:

do j = 1,n
do i = 1,n
a(i,j) = fcn(i,j)
end do
end do

The calculation of elements is independent of

one another - leads to an embarrassingly parallel
situation.
The problem should be computationally
intensive.

Array Processing
Parallel Solution 1

Arrays elements are distributed so that each

processor owns a portion of an array (subarray).
Independent calculation of array elements
ensures there is no need for communication
between tasks.
Distribution scheme is chosen by other criteria,
e.g. unit stride (stride of 1) through the
subarrays. Unit stride maximizes cache/memory
usage.
Since it is desirable to have unit stride through
the subarrays, the choice of a distribution
scheme depends on the programming language.
See the Block - Cyclic Distributions Diagram
for the options.
After the array is distributed, each task executes
the portion of the loop corresponding to the data

it owns. For example, with Fortran block

distribution:

do j = mystart, myend
do i = 1,n
a(i,j) = fcn(i,j)
end do
end do

Notice that only the outer loop variables are

different from the serial solution.

One Possible Solution:

Implement as a Single Program Multiple Data (SPMD) model.

Master process initializes array, sends info to worker processes and receives results.
Worker process receives info, performs its share of computation and sends results to
master.
Using the Fortran storage scheme, perform block distribution of the array.
Pseudo code solution: red highlights changes for parallelism.

Methodical Design of Parallel Algorithms

There is no simple recipe for designing parallel algorithms. However, it can
benefited from a methodological approach that maximizes the range of options, that
provides mechanisms for evaluating alternatives, and that reduces the cost of
backtracking from wrong choices. The design methodology allows the programmer to
focus on machine-independent issues such as concurrency in the early stage of
design process, and machine-special aspects of design are delayed until late in the
design process. As suggested by Ian Foster, this methodology organizes the design
process into four distinct stages [12]:
partitioning
communication
agglomeration
mapping
The first two stages seek to develop concurrent and scalable algorithms, and the last
two stages focus on locality and performance-related issues as summarized below:
Partitioning
It refers to decomposing of the computational activities and the data on which it
operates into several small tasks. The decomposition of the data associated with a
problem is known as domain/data decomposition, and the decomposition of

computation into disjoint tasks is known as functional decomposition.

Communication
It focuses on the flow of information and coordination among the tasks that are
created during the partitioning stage. The nature of the problem and the
decomposition method determine the communication pattern among these
cooperative tasks of a parallel program. The four popular communication patterns
commonly used in parallel programs are: local/global, structured/unstructured,
static/dynamic, and synchronous/asynchronous.
Agglomeration
In this stage, the tasks and communication structure defined in the first two stages are
evaluated in terms of performance requirements and implementation costs. If
required, tasks are grouped into larger tasks to improve performance or to reduce
development costs. Also, individual communications may be bundled into a super
communication. This will help in reducing communication costs by increasing
computation and communication granularity, gaining flexibility in terms of scalability
and mapping decisions, and reducing software-engineering costs.
Mapping
It is concerned with assigning each task to a processor such that it maximizes
utilization of system resources (such as CPU) while minimizing the communication
costs. Mapping decisions can be taken statically (at compile-time/before program
execution) or dynamically at runtime by load-balancing methods as discussed in
Volume 1 of this book [8] and Chapter 17.
Several grand challenging applications have been built using the above methodology (refer to Part III, Algorithms and Applications, for further details on the
development of real-life applications using the above methodology).
1.7 Parallel Programming Paradigms
It has been widely recognized that parallel applications can be classi_ed into some
well de_ned programming paradigms. A few programming paradigms are used repeatedly to develop many parallel programs. Each paradigm is a class of algorithms
that have the same control structure [19].
Experience to date suggests that there are a relatively small number of paradigms underlying most parallel programs [7]. The choice of paradigm is determined
by the available parallel computing resources and by the type of parallelism inherent in the problem. The computing resources may de_ne the level of granularity
that can be e_ciently supported on the system. The type of parallelism reects
Section 1.7 Parallel Programming Paradigms 17
the structure of either the application or the data and both types may exist in different parts of the same application. Parallelism arising from the structure of the
application is named as functional parallelism. In this case, di_erent parts of the
program can perform di_erent tasks in a concurrent and cooperative manner. But
parallelism may also be found in the structure of the data. This type of parallelism
allows the execution of parallel processes with identical operation but on di_erent

parts of the data.

1.7.1 Choice of Paradigms
Most of the typical distributed computing applications are based on the very popular client/server paradigm. In this paradigm, the processes usually communicate
through Remote Procedure Calls (RPCs), but there is no inherent parallelism in
this sort of applications. They are instead used to support distributed services, and
thus we do not consider this paradigm in the parallel computing area.
In the world of parallel computing there are several authors which present a
paradigm classi_cation. Not all of them propose exactly the same one, but we can
create a superset of the paradigms detected in parallel applications.
For instance, in [22], a theoretical classi_cation of parallel programs is presented
and broken into three classes of parallelism:
1. processor farms, which are based on replication of independent jobs;
2. geometric decomposition, based on the parallelisation of data structures; and
3. algorithmic parallelism, which results in the use of data ow.
Another classi_cation was presented in [19]. The author studied several parallel
applications and identi_ed the following set of paradigms:
1. pipelining and ring-based applications;
2. divide and conquer;
3. master/slave; and
4. cellular automata applications, which are based on data parallelism.
The author of [23] also proposed a very appropriate classi_cation. The problems
were divided into a few decomposition techniques, namely:
1. Geometric decomposition: the problem domain is broken up into smaller domains and each process executes the algorithm on each part of it.
2. Iterative decomposition: some applications are based on loop execution where
each iteration can be done in an independent way. This approach is implemented through a central queue of runnable tasks, and thus corresponds to
the task-farming paradigm.
18 Parallel Programming Models and Paradigms Chapter 1
3. Recursive decomposition: this strategy starts by breaking the original problem
into several subproblems and solving these in a parallel way. It clearly corresponds to a divide and conquer approach.
4. Speculative decomposition: some problems can use a speculative decomposition
approach: N solution techniques are tried simultaneously, and (N-1) of them
are thrown away as soon as the _rst one returns a plausible answer. In some
cases this could result optimistically in a shorter overall execution time.
5. Functional decomposition: the application is broken down into many distinct
phases, where each phase executes a di_erent algorithm within the same problem. The most used topology is the process pipelining.
In [15], a somewhat di_erent classi_cation was presented based on the temporal
structure of the problems. The applications were thus divided into:

1. synchronous problems, which correspond to regular computations on regular

data domains;
2. loosely synchronous problems, that are typi_ed by iterative calculations on
geometrically irregular data domains;
3. asynchronous problems, which are characterized by functional parallelism that
is irregular in space and time; and
4. embarrassingly parallel applications, which correspond to the independent
execution of disconnected components of the same program.
Synchronous and loosely synchronous problems present a somehow di_erent synchronization structure, but both rely on data decomposition techniques. According
to an extensive analysis of 84 real applications presented in [14], it was estimated
that these two classes of problems dominated scienti_c and engineering applications
being used in 76 percent of the applications. Asynchronous problems, which are
for instance represented by event-driven simulations, represented 10 percent of the
studied problems. Finally, embarrassingly parallel applications that correspond to
the master/slave model, accounted for 14 percent of the applications.
The most systematic de_nition of paradigms and application templates was presented in [6]. It describes a generic tuple of factors which fully characterizes a parallel algorithm including: process properties (structure, topology and execution),
interaction properties, and data properties (partitioning and placement). That classi_cation included most of the paradigms referred so far, albeit described in deeper
detail.
To summarize, the following paradigms are popularly used in parallel programming:
_ Task-Farming (or Master/Slave)
_ Single Program Multiple Data (SPMD)
Section 1.7 Parallel Programming Paradigms 19
_ Data Pipelining
_ Divide and Conquer
_ Speculative Parallelism
1.7.2 Task-Farming (or Master/Slave)
The task-farming paradigm consists of two entities: master and multiple slaves. The
master is responsible for decomposing the problem into small tasks (and distributes
these tasks among a farm of slave processes), as well as for gathering the partial
results in order to produce the _nal result of the computation. The slave processes
execute in a very simple cycle: get a message with the task, process the task, and
send the result to the master. Usually, the communication takes place only between
the master and the slaves.
Task-farming may either use static load-balancing or dynamic load-balancing.
In the _rst case, the distribution of tasks is all performed at the beginning of the
computation, which allows the master to participate in the computation after each
slave has been allocated a fraction of the work. The allocation of tasks can be done

once or in a cyclic way. Figure 1.4 presents a schematic representation of this _rst
approach.
The other way is to use a dynamically load-balanced master/slave paradigm,
which can be more suitable when the number of tasks exceeds the number of available processors, or when the number of tasks is unknown at the start of the application, or when the execution times are not predictable, or when we are dealing
with unbalanced problems. An important feature of dynamic load-balancing is the
ability of the application to adapt itself to changing conditions of the system, not
just the load of the processors, but also a possible recon_guration of the system
resources. Due to this characteristic, this paradigm can respond quite well to the
failure of some processors, which simpli_es the creation of robust applications that
are capable of surviving the loss of slaves or even the master.
At an extreme, this paradigm can also enclose some applications that are based
on a trivial decomposition approach: the sequential algorithm is executed simultaneously on di_erent processors but with di_erent data inputs. In such applications
there are no dependencies between di_erent runs so there is no need for communication or coordination between the processes.
This paradigm can achieve high computational speedups and an interesting degree of scalability. However, for a large number of processors the centralized control
of the master process can become a bottleneck to the applications. It is, however,
possible to enhance the scalability of the paradigm by extending the single master
to a set of masters, each of them controlling a di_erent group of process slaves.
1.7.3 Single-Program Multiple-Data (SPMD)
The SPMD paradigm is the most commonly used paradigm. Each process executes
basically the same piece of code but on a di_erent part of the data. This involves
20 Parallel Programming Models and Paradigms Chapter 1
Master
distribute tasks
Slave 1 Slave 2 Slave 3 Slave 4
Terminate
Collect
Results
communications
Figure 1.4 A static master/slave structure.
the splitting of application data among the available processors. This type of parallelism is also referred to as geometric parallelism, domain decomposition, or data
parallelism. Figure 1.5 presents a schematic representation of this paradigm.
Many physical problems have an underlying regular geometric structure, with
spatially limited interactions. This homogeneity allows the data to be distributed
uniformly across the processors, where each one will be responsible for a de_ned
spatial area. Processors communicate with neighbouring processors and the communication load will be proportional to the size of the boundary of the element,

while the computation load will be proportional to the volume of the element. It
may also be required to perform some global synchronization periodically among
all the processes. The communication pattern is usually highly structured and extremely predictable. The data may initially be self-generated by each process or
may be read from the disk during the initialization stage.
SPMD applications can be very e_cient if the data is well distributed by the
processes and the system is homogeneous. If the processes present di_erent workloads or capabilities, then the paradigm requires the support of some load-balancing
scheme able to adapt the data distribution layout during run-time execution.
This paradigm is highly sensitive to the loss of some process. Usually, the loss
Section 1.7 Parallel Programming Paradigms 21
of a single process is enough to cause a deadlock in the calculation in which none
of the processes can advance beyond a global synchronization point.
Distribute Data
Calculate Calculate Calculate Calculate
Exchange Exchange Exchange Exchange
Calculate Calculate Calculate Calculate
Collect Results
Figure 1.5 Basic structure of a SPMD program.
1.7.4 Data Pipelining
This is a more _ne-grained parallelism, which is based on a functional decomposition
approach: the tasks of the algorithm, which are capable of concurrent operation,
are identi_ed and each processor executes a small part of the total algorithm. The
pipeline is one of the simplest and most popular functional decomposition paradigms. Figure 1.6 presents the structure of this model.
Processes are organized in a pipeline { each process corresponds to a stage of the
pipeline and is responsible for a particular task. The communication pattern can
be very simple since the data ows between the adjacent stages of the pipeline. For
this reason, this type of parallelism is also sometimes referred to as data ow parallelism. The communication may be completely asynchronous. The e_ciency of this
paradigm is directly dependent on the ability to balance the load across the stages
of the pipeline. The robustness of this paradigm against recon_gurations of the
system can be achieved by providing multiple independent paths across the stages.
This paradigm is often used in data reduction or image processing applications.
1.7.5 Divide and Conquer
The divide and conquer approach is well known in sequential algorithm development. A problem is divided up into two or more subproblems. Each of these
subproblems is solved independently and their results are combined to give the _nal result. Often, the smaller problems are just smaller instances of the original
22 Parallel Programming Models and Paradigms Chapter 1
Process 1 Process 2 Process 3
Phase A Phase B Phase C

Input Output
Figure 1.6 Data pipeline structure.
problem, giving rise to a recursive solution. Processing may be required to divide
the original problem or to combine the results of the subproblems. In parallel divide
and conquer, the subproblems can be solved at the same time, given su_cient parallelism. The splitting and recombining process also makes use of some parallelism,
but these operations require some process communication. However, because the
subproblems are independent, no communication is necessary between processes
working on di_erent subproblems.
We can identify three generic computational operations for divide and conquer:
split, compute, and join. The application is organized in a sort of virtual tree: some
of the processes create subtasks and have to combine the results of those to produce
an aggregate result. The tasks are actually computed by the compute processes at
the leaf nodes of the virtual tree. Figure 1.7 presents this execution.
main problem
sub-problems
split
join
Figure 1.7 Divide and conquer as a virtual tree.
The task-farming paradigm can be seen as a slightly modi_ed, degenerated form
of divide and conquer; i.e., where problem decomposition is performed before tasks
are submitted, the split and join operations is only done by the master process and
Section 1.8 Programming Skeletons and Templates 23
all the other processes are only responsible for the computation.
In the divide and conquer model, tasks may be generated during runtime and
may be added to a single job queue on the manager processor or distributed through
several job queues across the system.
The programming paradigms can be mainly characterized by two factors: decomposition and distribution of the parallelism. For instance, in geometric parallelism both the decomposition and distribution are static. The same happens with
the functional decomposition and distribution of data pipelining. In task farming, the work is statically decomposed but dynamically distributed. Finally, in the
divide and conquer paradigm both decomposition and distribution are dynamic.
1.7.6 Speculative Parallelism
This paradigm is employed when it is quite di_cult to obtain parallelism through
any one of the previous paradigms. Some problems have complex data
dependencies,
which reduces the possibilities of exploiting the parallel execution. In these cases,
an appropriate solution is to execute the problem in small parts but use some
speculation or optimistic execution to facilitate the parallelism.
In some asynchronous problems, like discrete-event simulation [17], the system
will attempt the look-ahead execution of related activities in an optimistic assump-

tion that such concurrent executions do not violate the consistency of the problem
execution. Sometimes they do, and in such cases it is necessary to rollback to some
previous consistent state of the application.
Another use of this paradigm is to employ di_erent algorithms for the same
problem; the _rst one to give the _nal solution is the one that is chosen.
1.7.7 Hybrid Models
The boundaries between the paradigms can sometimes be fuzzy and, in some applications, there could be the need to mix elements of di_erent paradigms. Hybrid
methods that include more than one basic paradigm are usually observed in some
large-scale parallel applications. These are situations where it makes sense to mix
data and task parallelism simultaneously or in di_erent parts of the same program.