Different Processor Frameworks.


65 views
Uploaded on:
Category: Sports / Games
Description
Multicomputer Hardware (3) Network interface sheets in a multicomputer. Low-Level ... consecutive consistency (b) circulated framework may return old worth ...
Transcripts
Slide 1

Various Processor Systems 8.1 Multiprocessors 8.2 Multicomputers 8.3 Distributed frameworks Chapter 8

Slide 2

Multiprocessor Systems Continuous requirement for quicker PCs shared memory model message passing multiprocessor wide territory disseminated framework

Slide 3

Multiprocessors Definition: A PC framework in which two or more CPUs share full access to a typical RAM

Slide 4

Multiprocessor Hardware (1) Bus-based multiprocessors

Slide 5

Multiprocessor Hardware (2) UMA Multiprocessor utilizing a crossbar switch

Slide 6

Multiprocessor Hardware (3) UMA multiprocessors utilizing multistage exchanging systems can be worked from 2x2 switches (a) 2x2 switch (b) Message group

Slide 7

Multiprocessor Hardware (4) Omega Switching Network

Slide 8

Multiprocessor Hardware (5) NUMA Multiprocessor Characteristics Single location space noticeable to all CPUs Access to remote memory by means of summons LOAD STORE Access to remote memory slower than to neighborhood

Slide 9

Multiprocessor Hardware (6) (a) 256-hub registry based multiprocessor (b) Fields of 32-bit memory address (c) Directory at hub 36

Slide 10

Multiprocessor OS Types (1) Each CPU has its own working framework Bus

Slide 11

Multiprocessor OS Types (2) Master-Slave multiprocessors Bus

Slide 12

Multiprocessor OS Types (3) Symmetric Multiprocessors SMP multiprocessor model Bus

Slide 13

Multiprocessor Synchronization (1) TSL guideline can fall flat if transport as of now bolted

Slide 14

Multiprocessor Synchronization (2) Multiple locks used to evade reserve whipping

Slide 15

Multiprocessor Synchronization (3) Spinning versus Switching at times CPU must hold up holds up to gain prepared rundown In different cases a decision exists turning squanders CPU cycles exchanging goes through CPU cycles additionally conceivable to settle on partitioned choice every time bolted mutex experienced

Slide 16

Multiprocessor Scheduling (1) Timesharing note utilization of single information structure for booking

Slide 17

Multiprocessor Scheduling (2) Space sharing numerous strings at same time over numerous CPUs

Slide 18

Multiprocessor Scheduling (3) Problem with correspondence between two strings both have a place with procedure A both coming up short on stage

Slide 19

Multiprocessor Scheduling (4) Solution: Gang Scheduling Groups of related strings planned as a unit (a posse) All individuals from group run all the while on various timeshared CPUs All group individuals begin and end time cuts together

Slide 20

Multiprocessor Scheduling (5) Gang Scheduling

Slide 21

Multicomputers Definition: Tightly-coupled CPUs that don\'t share memory Also known as bunch PCs bunches of workstations (COWs)

Slide 22

Interconnection topologies (a) solitary switch (b) ring (c) lattice (d) twofold torus (e) 3D square (f) hypercube Multicomputer Hardware (1)

Slide 23

Multicomputer Hardware (2) Switching plan store-and-forward parcel exchanging

Slide 24

Multicomputer Hardware (3) Network interface loads up in a multicomputer

Slide 25

Low-Level Communication Software (1) If a few procedures running on hub need system access to send bundles … Map interface load up to all procedure that need it If piece needs access to organize … Use two system loads up one to client space, one to portion

Slide 26

Low-Level Communication Software (2) Node to Network Interface Communication Use send & get rings facilitates principle CPU with on-load up CPU

Slide 27

User Level Communication Software (a) Blocking send call Minimum administrations gave send and get charges These are blocking (synchronous) calls (b) Nonblocking send call

Slide 28

Remote Procedure Call (1) Steps in making a remote technique call the stubs are shaded dark

Slide 29

Remote Procedure Call (2) Implementation Issues Cannot pass pointers call by reference gets to be duplicate reestablish (however may come up short) Weakly wrote dialects customer stub can\'t decide size Not generally conceivable to decide parameter sorts Cannot utilize worldwide variables may get moved to remote machine

Slide 30

Distributed Shared Memory (1) Note layers where it can be actualized equipment working framework client level programming

Slide 31

Distributed Shared Memory (2) Replication (a) Pages circulated on 4 machines (b) CPU 0 peruses page 10 (c) CPU 1 peruses page 10

Slide 32

Distributed Shared Memory (3) False Sharing Must likewise accomplish consecutive consistency

Slide 33

Multicomputer Scheduling Load Balancing (1) Graph-theoretic deterministic calculation Process

Slide 34

Load Balancing (2) Sender-started dispersed heuristic calculation over-burden sender

Slide 35

Load Balancing (3) Receiver-started appropriated heuristic calculation under stacked recipient

Slide 36

Distributed Systems (1) Comparison of three sorts of numerous CPU frameworks

Slide 37

Distributed Systems (2) Achieving consistency with middleware

Slide 38

Network Hardware (1) Computer Ethernet (a) great Ethernet (b) exchanged Ethernet (a) (b)

Slide 39

Network Hardware (2) The Internet

Slide 40

Network Services and Protocols (1) Network Services

Slide 41

Network Services and Protocols (2) Internet Protocol Transmission Control Protocol Interaction of conventions

Slide 42

Document-Based Middleware (1) The Web a major coordinated diagram of archives

Slide 43

Document-Based Middleware (2) How the program gets a page Asks DNS for IP address DNS answers with IP address Browser makes association Sends ask for indicated page Server sends record TCP association discharged Browser shows content Browser brings, shows pictures

Slide 44

File System-Based Middleware (1) Transfer Models (a) transfer/download model (b) remote access model (b) (a)

Slide 45

File System-Based Middleware (2) Naming Transparency (b) Clients have same perspective of document framework (c) Alternatively, customers with various perspective

Slide 46

File System-Based Middleware (3) Semantics of File sharing (a) solitary processor gives successive consistency (b) conveyed framework may return old quality

Slide 47

File System-Based Middleware (4) AFS – Andrew File System workstations gathered into cells note position of venus and bad habit Client\'s perspective

Slide 48

Shared Object-Based Middleware (1) Main components of CORBA based framework Common Object Request Broker Architecture

Slide 49

Shared Object-Based Middleware (2) Scaling to huge frameworks recreated objects adaptability Globe intended to scale to a billion clients a trillion items around the globe

Slide 50

Shared Object-Based Middleware (3) Globe organized article

Slide 51

Shared Object-Based Middleware (4) A dispersed shared article in Globe can have its state replicated on numerous PCs without a moment\'s delay

Slide 52

Shared Object-Based Middleware (5) Internal structure of a Globe object

Slide 53

Coordination-Based Middleware (1) Linda free procedures impart by means of dynamic tuple space Tuple like a structure in C, record in Pascal Operations: out, in, read, eval

Slide 54

Coordination-Based Middleware (2) Publish-Subscribe engineering

Slide 55

Coordination-Based Middleware (3) Jini - in view of Linda model gadgets connected to a system offer, use administrations Jini Methods read compose take advise

Recommended
View more...