没有合适的资源?快使用搜索试试~ 我知道了~
首页动态系统模拟:探索建模与分布式仿真
动态系统模拟:探索建模与分布式仿真
需积分: 15 3 下载量 113 浏览量
更新于2024-07-18
收藏 2.46MB PDF 举报
"《Modelling and Simulation_ Exploring Dynamic System Behaviour》是一本专注于建模与模拟动态系统行为的教材,作者Richard M. Fujimoto博士来自乔治亚理工学院。书中包含大量的代码实例,旨在帮助读者理解和应用平行及分布式模拟系统。本书是Wiley Series on Parallel and Distributed Computing系列的一部分,由Albert Y. Zomaya编辑。"
在现代信息技术领域,建模和模拟是理解和预测复杂系统行为的重要工具。动态系统,如计算机网络、分布式计算环境、物流系统、金融市场等,其行为往往难以通过直觉或简单的数学分析来完全理解。此时,建模和模拟就显得尤为关键。
建模是将现实世界的问题转化为数学或逻辑形式的过程,它可以抽象出系统的关键特征并忽略不重要的细节。在动态系统中,建模通常涉及离散事件模拟(Discrete Event Simulation, DES)或连续时间模拟(Continuous Time Simulation)。离散事件模拟关注系统中的独立事件,而连续时间模拟则处理随着时间连续变化的状态。
分布式和并行模拟是处理大规模复杂系统的关键技术。Richard M. Fujimoto的书深入探讨了这些主题,讲解如何在并行和分布式环境中进行有效的模拟。并行模拟允许同时处理多个事件或任务,显著提高了模拟的速度和效率,特别是在处理大量数据和计算密集型问题时。分布式模拟则涉及多个计算节点协作完成一个模拟任务,这种技术在处理地理上分散或网络化系统时特别有用。
书中提供的代码实例对于学习和实践这些概念至关重要。通过实际操作,读者可以更好地掌握如何构建和运行模拟,以及如何分析和解释结果。这些实例可能涵盖多种编程语言,如C++、Java或Python,涵盖了各种模拟库和框架的使用。
此外,这本书还可能涵盖了版权和许可的信息,强调未经许可不得复制或分发内容,除非符合美国1976年版权法第107或108节的规定,或者通过支付适当的副本费用获得了授权。
《Modelling and Simulation_ Exploring Dynamic System Behaviour》为读者提供了一个深入理解动态系统行为的平台,通过并行和分布式模拟技术,使读者能够处理复杂的计算挑战,并在实践中提升自己的技能。无论是学生、研究人员还是工程专业人士,都能从这本书中获益匪浅。
10
BACKGROUND AND APPLICATIONS
1.4 APPLICATIONS
11
success
of
the SIMNET experiment has had far-reaching effects throughout the
defense modeling and simulation community in the United States. SIMNET was
replaced by what came to be known
as
Distributed Interactive Simulation (DIS)
where standards were defined
to
support interoperability among autonomous training
simulators in geographically distributed simulation environments.
A second major development springing from SIMNET was the Aggregate Level
Simulation Protocol (ALSP) work that applied the SIMNET concept
of
interoper-
ability
to
war game simulations. ALSP enabled war game simulations from the
Army, Air Force, and Navy, for example,
to
be brought together in a single exercise
to analyze joint military operations. ALSP used synchronization protocols discussed
earlier for analytic simulations; it represents perhaps the most extensive application
of
that technology
to
date. Work in the ALSP community proceeded concurrently
with work in the DIS community that was focused primarily (though not exclusively)
on training.
The next major milestone in the evolution
of
this technology was the develop-
ment
of
the High Level Architecture (HLA) which began in
1995
and resulted in the
so-called
baseline definition in August 1996. HLA
is
important for several reasons.
From a practical standpoint, HLA was mandated in September 1996
as
the standard
architecture for all modeling and simulation activities in the Department
of
Defense
in the United States. All DoD simulations are required to become
HLA compliant (or
obtain approval for an exception
to
this requirement) by 1999. From a technical
standpoint, HLA
is
important because it provides a single architecture that spans
both analytic and virtual environment simulations. In some respects it can be viewed
as
a merging
of
DIS and ALSP into a single architecture. Prior to the HLA effort,
work in the paralleljdistributed analytic simulation community and the distributed
virtual environment communities proceeded largely independent
of
each other. HLA
was a landmark effort in that it began integrating these technologies in a significant
way.
At the time
of
this writing, the initial baseline definition
of
the HLA and its
realization in prototype versions
of
the Runtime Infrastructure (RTI) have been
completed and standardization activities are in progress. Migration
of
DIS standards
to
the HLA
is
also under
way.
1.3.3 Interactive Gaming and Internet Communities
A second major thread
of
activity in distributed virtual environments for nonmilitary
applications grew from the interactive gaming and Internet communities. Just
as
defense simulations originated from "platform-level" simulators for tanks and
aircraft, nonmilitary
DVE
work originated in "immersive" games such
as
Adventure
and Dungeons and Dragons. Adventure was a fantasy computer game created at
Xerox Palo Alto Research Center (PARC) in California in the mid-1970s. In it a
user/player explored a rich computer-generated fantasy world, most
of
which was
underground in a maze
of
caves and hidden passage ways. Adventure was a text-
based game where users typed short phrases
to
describe their actions (for example,
"move up"), and were given word descriptions
of
objects and rooms they
encountered in their journey. This fantasy world was complete with a rich variety
of
hidden treasures and a wide assortment
of
other computer-generated creatures,
both friend and foe, that could help or hinder the player from finding and obtaining
the treasures. A skilled player could slay harmful adversaries such
as
dragons with
various weapons, such
as
swords and magic potions giving the partaker special
abilities for a limited amount
of
time. These weapons could be found in different
areas
of
the virtual world. Adventure was developed in the 1970s and 1980s before
powerful personal computers were widely available.
Yet
despite the limited interac-
tion allowed by a text-oriented program, it was a very popular game among the
college students who were lucky enough
to
have computer access. Computer and
video games
of
this nature continue to thrive today, greatly enhanced with audio
effects and computer graphics.
Adventure was a
single-player game. A second, key ingredient in the develop-
ment
of
DVEs was the introduction
of
multiple players
to
the virtual world. Though
not initially computerized, the popular game
of
Dungeons and Dragons, also from
the mid-1970s,
is
credited with being the catalyst for this development. This was a
pencil and paper role-playing game where players gathered to play out roles
as
knights and sorcerers in a made-up world created by one
of
the players, referred to
as
the dungeon master. The actual environment could be
as
simple
as
a written
description
of
the various portions
of
the virtual world, or
as
elaborate
as
scale
models.
Computer-generated fantasy games and multiple users/players came together in
the early 1980s with the MultiUser Dungeon (MUD) game developed at the
University
of
Essex in England.
Today,
the term MUD
is
associated with multiplayer
games
of
this sort in general,
as
opposed
to
any particular game. Further the
applications for DVEs extend far beyond games, and a substantial amount
of
work
has been geared toward nongaming applications.
In addition
to
computer-generated virtual worlds and the inclusion
of
multiple
players, a third critical ingredient in the development
of
DVEs was the unprece-
dented expansion and growth
of
the worldwide network
of
computer networks
known
as
the Internet. With the Internet a virtual environment can support multiple
users who may be scattered around the globe. Multiplayer games with geographi-
cally distributed players are flourishing in the 1990s, despite limited bandwidth (for
example modem lines with
as
little
as
9600 bits per second) and relatively high
network latencies. Continued increases in modem bandwidth (at the time
of
this
writing in the late 1990s,
50
Kbits/second modems are becoming widely available)
and megabit per second bandwidths (for example, via cable modems or other
technologies) are on their
way.
The communication bottlenecks that have hitherto
restricted widespread use
of
distributed virtual environments may be a thing
of
the
past.
1.4 APPLICATIONS
With the above historical context,
we
now survey some
of
the applications where
parallel and distributed simulation technologies have been applied. While far from
12
BACKGROUND AND APPLICATIONS
1.4 APPLICATIONS
13
being complete, this list gives a flavor
of
some
of
the current and potential uses
of
the technology.
1.4.1
Military Applications
It
is
clear that the military establishment has had a major role in developing
distributed simulation technology for virtual environments, and to a lesser though
still significant extent, parallel simulation technology for analytic simulation
applications. Some
of
the most prominent military applications utilizing this
technology are
as
follows:
1.
War
gaming simulations. These simulations are often used
to
evaluate different
strategies for attacking or defending against an opposing force, or for
acquisition decisions to determine the number and type
of
weapon systems
that should be purchased
to
be prepared for future engagements. The
simulation
is
typically composed
of
models for battalions, divisions, and
so
forth. Because these simulations usually model groups
of
units rather than
individual platforms (for example, aircraft and tanks), they are sometimes also
referred to
as
aggregated simulations.
Two
noteworthy examples
of
the
application
of
parallel discrete event simulation techniques
to
war game
simulations are the
Concurrent Theater Level Simulation (CTLS) (Wieland,
Hawley et
al.
1989) and Aggregate Level Simulation Protocol (ALSP) (Wilson
and Weatherly 1994) discussed earlier. The underlying execution mechanism
for CTLS was a parallel simulation executive using a synchronization
algorithm called
Time
Warp.
ALSP used another algorithm called the
ChandyjMisrajBryant null message protocol. These algorithms will be
discussed in detail in Chapters 4 and
3,
respectively.
2.
Training environments. As discussed earlier, these simulations embed pilots,
tank operators, commanding officers and their staffs, and the like, into an
environment
to
train personnel for actual combat. In contrast
to
aggregated
simulations, many training environments use platform-level simulations that
do
model individual tanks, aircraft, and
so
forth.
3.
Test
and evaluation (T&E). While training simulations embed humans into a
synthetic battlefield, T&E simulations embed physical components (for
example, a new sensor
for
detecting missile launches) into a virtual environ-
ment, often
to
evaluate the effectiveness
of
proposed new devices or to verify
that manufactured devices operate at reported specifications. The T
&E
simulation community has sometimes been referred to
as
the "Consumer
Reports" for the military because they evaluate new products before they are
manufactured and eventually deployed.
As discussed earlier, the High Level Architecture effort attempts
to
integrate
simulations from these three domains in order to facilitate reuse
of
simulation
models in new contexts, thereby reducing the cost
of
developing new simulators.
1.4.2 Entertainment
The number
of
real F-15 pilots that can benefit from immersion into a computer-
generated dogfight
is
dwarfed by the number
of
wannabe pilots that are looking for
recreation on a Saturday night. Application
of
distributed simulation technology to
the entertainment industry will (and already is) leading
to
the most significant impact
of
this technology on the average citizen. Single-player video arcade games cannot
provide the same kind
of
entertainment
as
interactively competing with friends (or
strangers) in a computer-generated virtual world.
Distributed simulation technology can be applied in amusement park and arcade
centers where players are co-located but interact with each other and computer-
generated entities over a local area network. These systems sometimes use costly
custom-designed hardware that can only be justified economically by repeated use
by
many users. Another emerging market
is
the multi-user home entertainment
industry where video game machines or personal computers are interconnected
through the Internet.
Entertainment and training systems employing distributed simulation technolo-
gies have much in common, but they also differ in many important respects.
Obviously, entertainment systems must be engaging. Unlike training simulators,
one does not have a captive audience where players must return to fulfill job
requirements. Thus the "realism"
of
the virtual environment may take second place
to
pure excitement and artistic effects. Economic factors
playa
much more dominant
role in the design
of
entertainment systems, sometimes requiring compromises that
would not be necessary in a multimillion dollar training system. Interoperability
among separately developed simulations
is
a fundamental goal in DIS, but it may be
viewed
as
undesirable
by
some in the entertainment industry. This would be true, for
example, when a company marketing proprietary entertainment systems has control,
as
a single vendor, over the simulations that will be included in the system.
1.4.3 Social Interactions
and
Business Collaborations
Another potentially far-reaching impact
of
distributed virtual environments IS m
creating new means for people
to
interact socially on the Internet. The Internet has
already made fundamental changes in the way people interact both in the office and
at
home. Many believe DVEs represent the next logical step in electronic social
interactions. Already users around the world can "meet" without ever leaving their
own home through Internet newsgroups and "chat rooms." Beyond this, a DVE
application can create more realistic social settings such
as
the one known
as
Diamond Park developed
by
Mitsubishi Electric Corp's MERL research laboratory.
Diamond Park provides a virtual park atmosphere where users can meet and interact
in various settings such
as
the park's cafe, walkways, or meeting areas (Waters and
Barrus 1997). Users can navigate through the park on foot or on bicycle and can
even race against each other! Virtual environments like this may be the norm in the
future for social interaction via the Internet.
14 BACKGROUND AND APPLICATIONS
Virtual environments can also provide a new means for interactions in the
business
world
between colleagues and clients. Entire "virtual corporations"
could be
created,
composed
of
employees who are based physically at different
locations or different companies but who are working together on ajoint venture. For
example,
one
can envision building designers and engineers at different locations
walking through a virtual design
of
a product (a building)
to
discuss and evaluate
design changes.
1.4.4
Education
and
Training
Nonmilitary applications for DVEs in education and training abound. Much work
has been accomplished in the medical community using virtual environments for
training
as
well
as
treatment
of
patients. Computer-generated environments can
provide a
more
cost-effective (and safe!) means for doctors
to
practice surgical
techniques. Experimental studies have been performed/conducted using virtual
environments
to
treat patients with various phobias such
as
a fear
of
heights.
Patients are exposed gradually and in a controlled way
to
(virtual) situations that
cause them
anxiety.
While much
of
the work
to
date in these areas has been focused
on single-user virtual environments (i.e.,
not distributed), extensions to DVEs to
allow for
users
to
remain in different geographic locations are clear. Work has also
focused on using
DVE
technology developed under DIS for nonmilitary applica-
tions, such
as
training air traffic controllers, or performing exercise drills for
emergency procedures, such
as
recovery from earthquakes or major accidents.
1.4.5 Telecommunication Networks
Analytic simUlations have long been used in the telecommunications industry
to
evaluate networking hardware, software, protocols, and services. The widespread
deployment
of
fiber optics technology has had important impacts on the use
of
simulation
in
modeling networks. First, this technology has brought about increased
use
of
telecommunication networks for applications other than voice communica-
tions, namely transmission
of
still images, data, and video. So-called Broadband
Integrated Services Digital Networks (B-ISDN) provide a single networking infra-
structure
to
carry these diverse types
of
traffic. Network designers have had
to
totally
rethink their designs, and tum toward simulation tools
to
aid them. Networking
technologies such
as
Asynchronous Transfer Mode or
ATM
2
have emerged
to
meet
the challenge
of
supporting these diverse types
of
traffic on a single network
infrastructure, (for example, see Partridge 1993).
Second, because the underlying network
is
based on fiber optic links that can
carry orders
of
magnitude more traffic than copper cables, simulations become more
time-consuming. This
is
because one must often model the network for at least the
duration
of
a conversation
to
collect useful data; that is, simulations
of
minutes to
2
An
unfortunate acronynm, this technology has nothing to do with Automated Teller Machines used in the
baking industry.
1.4 APPLICA
nONS
15
hours
of
network operation are required. Because B-ISDN networks carry orders
of
magnitude more traffic during this time period, the computation time
to
complete the
simulation increases
in
proportion. Parallel simulation techniques offer one approach
toward alleviating this problem.
A typical problem in telecommunications that calls out for
t~e
use
of
parallel
simulation
is
that
of
analyzing cell losses in
ATM
networks. A cell
IS
a 53-byte bl?ck
f data that
is
the basic unit transmitted through an
ATM
network. Each
ATM
SWitch
oontains buffers
to
hold cells waiting
to
be transmitted on links.
If
a link becomes
~ongested
and the buffers become full, subsequent cells that
r~quire
u~e
~f
that link
are
discarded. The cell loss
pro~ability
is
an important metrIc
th.a~
.mdicates
h~:,
frequently this happens.
ATM
SWitches
often target cell loss prob.a?IhtIes t? be 1
~
,
that is, only one in
10
9
cells
is
lost under anticipated traffic
.co.nditIOns.
ThiS
reqUIres
simulation
of
at least
1011
cell arrivals
to
obtain reliable statistIcal data. As
of
the late
1990s fast sequential simulation will execute on the order
of
10
5
events per second,
so
(o;timistically) equating simulation
of
ea~h
cell
a~rival
with
~
single
e.ven~,
such a
simulation will require more than
11
days, Just to Simulate a smgle
SWitch.
A second major area where parallel simulation may have a
signifi~ant
impact
is
in
the simulation
of
large networks such
as
the Internet. Here, high-performance
simulation engines are required because
of
the large number
of
entities that must
be simulated. Simulations
of
millions
of
mobile subscribers are sometimes needed.
1.4.6 Digital Logic Circuits
and
Computer Systems
Like telecommunication networks, simulations
of
digital electronic circuits and
computer systems
is
a second area where parallel simulation can play a
signific~t
role. Fast simulation
of
logic circuits
is
of
considerable interest
to
the electrolllc
computer-aided-design community because simulation
is
a major
~ottleneck
in
~he
design cycle. Final verification
of
a computer system may
reqUIre
weeks usmg
conventional sequential simulation techniques.
Much
of
the work in applying parallel simulation techniques
to
logic circuits has
been focused on the VHDL hardware description language that has become widely
used in industry. Several prototype parallel simulation systems have
bee~
develo~ed
that execute VHDL programs on multiple processor computers,
With
varymg
degrees
of
success reported in the literature. Successful demonstrations typically
report up
to
an order
of
magnitude reduction in execution
ti~e..
..
.,
While so-called gate-level logic simulations focus on modelmg mdividual
CirCUIts
for implementing primitive Boolean functions and storage elements, h.igher-level
simulations
of
computers using models for switches, processors, memOnes, and
so
forth are also used extensively in preliminary investigations
of
design alternatives.
Thes~
higher-level simulations often include simulated
ex~c~tions
of
bench~ark
programs on the modeled machine to evaluate it under :eahstIc
wor~loads.
Direct
execution
is
a technique where the benchmark program
IS
executed directly on the
machine used
to
perform the simulation rather than use a
(m~ch
slower) software
interpreter. Prototype parallel simulation systems
.using
.technlques
sU~h.
as
these
have been demonstrated
to
yield very accurate SimulatIon results, withm a
few
16 BACKGROUND AND APPLICATIONS
1.6 HARDWARE PLATFORMS
17
perce~t
of
measurements from a realized system, while delivering up to an order
of
magllltude reduction in computation time.
1.6 HARDWARE PLATFORMS
The hardware platforms
of
interest here contain a potentially large number
of
processors interconnected through a communication network. In most cases the
processor is a general purpose CPU (central processing unit), often identical to those
commonly found in personal computers and engineering workstations. The switch-
ing network may be as specific
as
a customized switch for a particular multi-
processor system, or as general as the Internet.
1.6.1 Parallel versus Distributed Computers
Multiple-CPU hardware platforms can be broadly classified into two categories:
parallel and distributed computers. Differences between these platforms are summar-
ized in Table 1.2. Parallel and distributed computing platforms are distinguished by
the physical area occupied by the computer. The processors in
parallel computers
are in close physical proximity, usually within a single cabinet, or a small number
of
adjacent cabinets in a machine room. These are usually homogeneous machines,
using processors from a single manufacturer. These machines normally provide
switching hardware tailored to the parallel computer, so the delay in transmitting a
message from one computer to another (referred to as the
communication latency) is
relatively
low.
This latency is typically a few microseconds to tens
of
microseconds
for a message containing a few bytes in contemporary machines. Latency
is
important because it has a large impact
on
performance;
if
latencies are large, the
computers may spend much
of
their time waiting for messages to be delivered. Here,
communication latency is perhaps the single most important technical aspect
differentiating parallel and distributed computers. There are three principal classes
of
parallel computers that are in use today: shared-memory multiprocessors,
distributed memory multicomputers,
and SIMD machines (see Fig. 1.1),
as
will be
elaborated upon momentarily.
Distributed computers cover a much broader geographic area. Their extent may
be confined to a single building, or may be as broad as across an entire nation or
even the world. Unlike parallel computers, each node
of
a distributed computer is
usually a stand-alone machine that includes its own memory and
I/O
devices.
Commercial off-the-shelf personal computers or engineering workstations, often
from different manufacturers, are usually used. Communication latencies are usually
TABLE 1.2 Contrasting parallel
and
distributed computers
1.4.7 Transportation
Simulation can play an important role in designing and managing road and air
transportation systems.
It
can be used as an analysis tool to evaluate the effectiveness
of
adding a new runway to an airport, or rerouting vehicular traffic after the
completion
of
a major sporting event. As alluded to earlier, it may be used "on-line"
in developing strategies to respond to an unexpected event, for example, congestion
resulting from adverse weather conditions.
1.5 UNDERLYING TECHNOLOGIES
Parallel and distributed simulation is made possible by the confluence
of
three
essential, underlying technologies:
• Integrated circuits. The first key ingredient is an inexpensive computer, thereby
making systems composed
of
tens, hundreds, or even thousands
of
computers
economically feasible. Fundamental to this development are steadily decreas-
ing costs
of
integrated circuits, driven largely by an increasing ability to
squeeze more and more circuits onto a single silicon chip. For example, the
cost
of
random access memory (RAM) that accounts for a significant portion
of
the cost
of
a personal computer or workstation has, over the long term,
decreased by 40% per year (Hennessy and Patterson 1996).
• High-speed intercomputer communications. There are two flavors
of
technol-
ogy at work here. On the one hand, high-speed switches enable one to
construct systems containing tens to hundreds or even thousands
of
processors
that reside
within a single cabinet or computer room. On the other hand,
advances in fiber optics technology is fueling a revolution in the telecommu-
nications industry, making possible computing systems
distributed across
continents.
These advances enable one to consider developing computer
applications utilizing many geographically distributed machines.
• Modeling and simulation. The final ingredient are technologies to enable
construction
of
models
of
actual or envisioned real-world systems that can (1)
be represented in the internal storage
of
a computer, and (2) be manipulated by
computer programs to emulate the evolution
of
the actual system over time.
Here, we are primarily concerned with
discrete event simulation, where
changes in the state
of
the simulation are viewed
as
occurring at distinct
points in time.
. In many applications other technologies such as graphics, human-computer
mterfaces, and databases clearly play a critical role. However, these technologies
are somewhat tangential to the focus
ofthis
book, and are not discussed further here.
Physical extent
Processors
Communication network
Communication latency
Parallel Computers
Machine room
Homogeneous
Customized switch
Less than 100 microseconds
Distributed Computers
Single building to global
Often heterogeneous
Commercial
LAN
or
WAN
Hundreds
of
microseconds
to seconds
18
BACKGROUND AND APPLICATIONS
1.6 HARDWARE PLATFORMS
19
Hardware Platforms
D;"lb"l
CO"P"""
networked
workstations
CPU
cache
• • •
CPU
cache
CPU
cache
Simulations executing on distributed computers are referred to
as
distributed
simulations.
Distributed simulations may be used for analytic purposes, or more
commonly for constructing distributed virtual environments. The latter
is
perhaps
the more common application for distributed simulation technology, and the term
distributed simulation in the literature sometimes refers exclusively
to
distributed
virtual environments.
1.6.2 Shared-Memory Multiprocessors
Shared-memory multiprocessors, distributed memory multicomputers, and SIMD
machines provide different programming models
to
the application. The distinguish-
ing property
of
the programming model for shared-memory multiprocessors
is
one
may
define variables that are accessible by different processors. Thus one can define
a variable
X that can be autonomously read or modified by one processor without the
intervention
of
another.
Shared variables and message passing are the two dominant forms
of
interpro-
cessor communications used in parallel and distributed simulation. Message-passing
mechanisms, widely used in parallel simulation, can be implemented using shared
memory by defining shared data structures (queues) to hold incoming or outgoing
messages.
One type
of
shared-memory machine, the symmetric multiprocessor (SMP), has
become increasingly popular. A typical shared-memory machine
is
depicted in
Figure
1.2.
These systems consist
of
off-the-shelf microprocessors connected to
memory through a high-speed switch, such
as
a bus. Frequently accessed instruc-
tions and data are stored in a high-speed
cache memory that
is
attached
to
each
processor. Typically the multiprocessor hardware automatically moves data and
instructions between the cache and "main" memories, so the programmer need not
be concerned with its operation, except perhaps
to
tune the program to maximize
performance. Consistency protocols are required to ensure that multiple copies
of
any shared variable residing in different caches remain up-to-date
if
one copy
is
modified; this
is
often realized in the hardware by either invalidating or updating
copies residing in other caches when one copy
is
changed. The Sun Enterprise
system
is
an example
of
a contemporary
SMP.
Personal computers (PCs) containing
Figure 1.2 Block diagram
of
a typical shared-memory multiprocessor.
Parallel Computers
shar~IMD
me~~;y
I
m~chines
distributed
memory
(multicomputers)
Figure 1.1 Taxonomy
of
important classes
of
parallel and distributed computers.
on the order
of
a
few
hundreds
of
microseconds for distributed computers with
processors in close proximity (for example, a single building), but they may
be
as
large
as
hundreds
of
milliseconds or even seconds for machines covering large
geographical areas. In the latter case, satellites may be used for some communication
links, contributing
to
increased latency. The latency
of
distributed computers
is
much
higher than parallel machines because
(l)
signals must traverse large physical
distances and (2) complex software protocols designed for interconnecting auton-
omous computers from different manufacturers are usually used rather than
customized hardware and software designed for a specific interconnection scheme.
While technological advances may be able to substantially reduce software over-
heads for communication, latency between geographically distributed machines
is
fundamentally limited by the speed
of
light, which
is
approximately
2.1
x
10
8
meters
per second in optical fiber, or 210 kilometers (
131
miles) per millisecond. A modem
microprocessor such
as
that included in personal computers can execute tens
to
hundreds
of
thousands
of
machine instructions (where each instruction can perform
a simple operation such
as
an integer addition) in one millisecond, so this
is
a
substantial amount
of
time from a computing perspective.
Recently the distinction between parallel and distributed computers has become
blurred with the advent
of
the network
of
workstations, which
is
a cluster
of
workstations interconnected through a high-speed switch usually confined
to
a single
room. Through the use
of
new switching techniques that bypass traditional
communication protocols, communication latency
of
these machines approach that
of
conventional parallel computers. Are these parallel or distributed computers?
Here, because
of
the close physical proximity
of
the machines, they are characterized
as
parallel computers, though often these systems are classified
as
distributed
machines. Because these machines often include aspects common
to
both parallel
and distributed computers, the characterization
is
perhaps not
so
important.
Simulations that execute on shared memory multiprocessors, multicomputers, or
SIMD machines are referred
to
as
parallel simulation programs. The focus
of
this
book with respect
to
parallel simulations
is
on discrete event simulations (discussed
in Chapter 2) used for analysis. The field concerned with this subject
is
called
parallel discrete event simulation (PDES). Here, the terms parallel simulation and
parallel discrete event simulation will be used synonymously.
剩余161页未读,继续阅读
2019-03-09 上传
2024-09-28 上传
2022-07-14 上传
2022-07-14 上传
2018-01-30 上传
2021-02-07 上传
2019-03-30 上传
2010-02-26 上传
2021-05-27 上传
weixin_39516685
- 粉丝: 0
- 资源: 43
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- MATLAB新功能:Multi-frame ViewRGB制作彩色图阴影
- XKCD Substitutions 3-crx插件:创新的网页文字替换工具
- Python实现8位等离子效果开源项目plasma.py解读
- 维护商店移动应用:基于PhoneGap的移动API应用
- Laravel-Admin的Redis Manager扩展使用教程
- Jekyll代理主题使用指南及文件结构解析
- cPanel中PHP多版本插件的安装与配置指南
- 深入探讨React和Typescript在Alias kopio游戏中的应用
- node.js OSC服务器实现:Gibber消息转换技术解析
- 体验最新升级版的mdbootstrap pro 6.1.0组件库
- 超市盘点过机系统实现与delphi应用
- Boogle: 探索 Python 编程的 Boggle 仿制品
- C++实现的Physics2D简易2D物理模拟
- 傅里叶级数在分数阶微分积分计算中的应用与实现
- Windows Phone与PhoneGap应用隔离存储文件访问方法
- iso8601-interval-recurrence:掌握ISO8601日期范围与重复间隔检查
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功