We accept

Definitions of Multiprocessors in Computing

A multiprocessor can be defined as the computer which uses several processing models under the involved control. Multi-processing is also thought as just how of using two or more than two CPUs within an individual computer. As everybody knows that we now have processors inside the computer systems, the multi processors, as the name suggests, be capable of support more than one cpu at a same time. Usually in multi-processing the processors are planned in the parallel form and therefore a large variety of the executions can be helped bring at the same time i. e. multi-processing assists with performing the same instructions a number of energy at a specific time. Some other related explanation of the multi processors are that multi-processing is the posting of the execution process by the interconnection of more than one microprocessor using securely or loosely lovers technology. Usually multi-processing tasks bears two simultaneous steps. You are the performing the duty of editing and enhancing and the other is the managing the data handling. A multi-processor device comprising, over a single semiconductor chip a plurality of processors including a first group of processors another band of processors; an initial bus to that your first band of processors is combined; a second bus to which the second band of processors is coupled; a first exterior bus user interface to which the first bus is coupled; and a second external bus software to which the second bus is coupled. The term multiprocessing is also used to make reference to a pc that has many impartial processing elements. The control elements are almost full computer systems in their own right. The primary difference is they have been freed from the encumbrance of communication with peripherals.


The processors are usually made up of the tiny and medium scale ICs which often includes a less or large numbers of the transistors. The multi processors consists of a computer architecture Most popular multiprocessor systems today use an SMP structures. Regarding multi-core processors, the SMP structures pertains to the cores, treating them as different processors. SMP systems allow any processor to focus on any task no matter where the data to the task are situated in recollection; with proper operating system support, SMP systems can easily move jobs between processors to balance the workload efficiently.


  • Increased processing power
  • Scale reference use to program requirements

Additional operating system responsibilities

  • All processors stay busy
  • Even distribution of procedures throughout the system
  • All processors work on steady copies of shared data
  • Execution of related processes synchronized
  • Mutual exclusion enforced

Multiprocessing is a type of processing where two or more processors work together to process several program simultaneously. Multi cpu systems have more than one processor that's why known as multi processor systems.

In multiprocessor system there is one master cpu and other are the Slave. If one cpu fails then get good at can assign the duty to other slave processor chip. But if Get good at will be fail than whole system will fail. Central part of Multiprocessor is the Get good at. All of them show the hard disk drive and Memory and other recollection devices.

Examples of multiprocessors

1. Quad-Processor Pentium Pro

  • SMP, bus interconnection.
  • 4 x 200 MHz Intel Pentium Pro processors.
  • 8 + 8 Kb L1 cache per cpu.
  • 512 Kb L2 cache per processor chip.
  • Snoopy cache coherence.
  • Compaq, HP, IBM, NetPower.
  • Windows NT, Solaris, Linux, etc.

2. SGI Origin 2000

  • NUMA, hypercube interconnection.
  • Up to 128 (64 x 2) MIPS R 10000 processors.
  • 32 + 32 Kb L1 cache per processor.
  • 4 Mb L2 cache per cpu.
  • Distributed directory-based cache coherence.
  • Automatic site migration/replication.
  • SGI IRIX with Pthreads

Classifications of multiprocessor architecture

  1. Nature of data path
  2. Interconnection scheme
  3. How processors share resources

Message-Passing Architectures

  • Separate address space for every single processor.
  • Processors communicate via meaning passing.

B) Shared-Memory Architectures

  • Single address space shared by all processors.
  • Processors communicate by memory space read/write.
  • SMP or NUMA.
  • Cache coherence is important concern.

1. Classifying Sequential and Parallel Architectures(DATA Way)

  • Stream: collection of bytes
  • Data stream
  • Instruction stream
  • Flynn's classifications:

MISD multiprocessing: MISD multiprocessing offers mainly the advantage of redundancy, since multiple processing items perform the same duties on a single data, reducing the chances of wrong results if one of the items fails. MISD architectures may require comparisons between handling units to detect failures. In addition to the redundant and fail-safe character of this type of multiprocessing, it includes few advantages, and it is very expensive. It generally does not improve performance. It can be implemented in a manner that is clear to software. It is employed inarray processorsand is implemented in problem tolerant machines.

MIMD multiprocessing: MIMD multiprocessing architecture is suitable for a wide variety of tasks where completely independent and parallel execution of instructions touching different models of data can be placed to fruitful use. For this reason, and because it is not hard to use, MIMD predominates in multiprocessing.

Processing is split into multiplethreads, each using its own hardware processor chip state, within an individual software-defined process or within multiple functions. Insofar as a system has multiple threads awaiting dispatch (either system or end user threads), this structures makes good use of hardware resources.

MIMD does increase issues of deadlock and source of information contention, however, since threads may collide in their usage of resources within an unstable way that is difficult to manage proficiently. MIMD requires special coding in the operating-system of the computer but will not require program changes unless the programs themselves use multiple threads (MIMD is translucent to single-threaded programs under most operating systems, if the programs do not voluntarily relinquish control to the Operating-system). Both system and individual software might need to use software constructs such assemaphores(also called locksorgates) to avoid one thread from interfering with another if indeed they should happen to cross pathways in referencing the same data. This gating or locking process raises code complexity, decreases performance, and greatly increases the amount of screening required, but not usually enough to negate the benefits of multiprocessing.

Similar conflicts can come up at the hardware level between processors (cache contention and problem, for example), and must usually be settled in hardware, or with a blend of software and hardware (e. g. , cache-clear instructions).

SISD multiprocessing: In asingle teaching stream, solo data streamcomputer one cpu sequentially processes instructions, each education functions one data item.

SIMD multiprocessing: In asingle instructions stream, multiple data streamcomputer one cpu handles a stream of instructions, each one of which can perform computations in parallel on multiple data locations. SIMD multiprocessing is suitable toparallel or vector control, in which a very large group of data can be split into parts that are individually subjected to identical but independent procedures. A single education stream directs the operation of multiple control units to execute the same manipulations simultaneously on potentially large amounts of data. For certain types of processing applications, this type of architecture can produce great increases in performance, in conditions of the elapsed time required to complete a given task. However, a downside to this architecture is that a large area of the system falls idle when programs or system tasks are carried out that cannot be divided into products that may be refined in parallel.

2. Interconnection scheme

Describes the way the system's components, such as processors and ram modules, are connected

  • Consists of nodes (components or switches) and links (cable connections)
  • Parameters used to evaluate interconnection schemes
  • Node degree
  • Bisection width
  • Network diameter
  • Cost of the interconnection scheme
  • Shared bus
  • Single communication route between all nodes
  • Contention can build up for shared bus
  • Fast for small multiprocessors
  • Form supernodes by hooking up several components with a distributed bus; use a far more scalable interconnection plan to connect supernodes
  • Dual-processor Intel Pentium

Shared bus multiprocessor firm.

  • Crossbar-switch matrix
  • Separate path out of every processor to every recollection module (or out of every to almost every other node when nodes contain both processors and storage modules)
  • High mistake tolerance, performance and cost
  • Sun UltraSPARC-III

Crossbar-s witch matrix multiprocessor business.

  • Hypercube
  • n -dimensional hypercube has 2 nodes in which each node is n connected to n neighbor nodes
  • Faster, more fault tolerant, but more expensive when compared to a 2-D mesh network
  • n CUBE (up to 8192 processors)
  • Multistage network
  • Switch nodes act as hubs routing communications between nodes
  • Cheaper, less problem tolerant, worse performance compared to a crossbar-switch matrix


  • Tightly coupled systems
  1. Processors share most resources including memory
  2. Communicate over shared buses using distributed physical memory
  3. Tasks and/or processors speak in an extremely synchronized fashion
  4. Communicates through the common shared memory
  5. Shared storage area system
  • Loosely coupled systems
  1. Processors do not share most resources
  2. Most communication through explicit emails or shared virtual memory (but not shared physical memory)
  3. Tasks or processors do not speak in a synchronized fashion
  4. Communicates by subject matter passing packets
  5. Overhead for data exchange is high
  6. Distributed storage area system

Comparison between them

  1. Loosely combined systems: more versatile, fault tolerant, scalable
  2. Tightly combined systems: more efficient, less burden to operating system programmers

Multiprocessor Operating System Organizations

Classify systems based how processors share operating-system responsibilities


  1. Master/slave
  2. Separate kernels
  3. Symmetrical organization

1) Master/slave organization

  • Master cpu executes the operating system
  • Slaves perform only consumer processors
  • Hardware asymmetry
  • Low mistake tolerance
  • Good for computationally intensive jobs

2) Separate kernels organization

  • Each processor executes its operating system
  • Some globally distributed operating-system data
  • Loosely coupled
  • Catastrophic failure improbable, but failure of one processor ends in termination of processes on that processor
  • Little contention over resources

Example: Tandem system

3) Symmetrical organization

  • Operating system handles a pool of similar processors
  • High amount of source sharing
  • Need for shared exclusion
  • Highest amount of problem tolerance of any organization
  • Some contention for resources

Example: BBN Butterfly

Memory Gain access to Architectures

  • Can classify multiprocessors based mostly about how processors talk about memory
  • Goal: Fast storage access from all processors to all memory
  • Contention in large systems makes this impractical

1) Uniform storage gain access to (UMA) multiprocessor

  • All processors share all memory
  • Access to any storage area page is nearly the same for any processors and everything ram modules (disregarding cache hits)
  • Typically uses shared bus or crossbar-switch matrix
  • Also called symmetric multiprocessing (SMP)
  • Small multiprocessors (typically two to eight processors)

2) Nonuniform ram access (NUMA) multiprocessor

  • Each node consists of a few processors and some of system memory space, which is local to that node
  • Access to local memory space faster than usage of global storage area (break of storage)
  • More scalable than UMA (fewer bus collisions)

3) Cache-only ram structures (COMA) multiprocessor

  • Physically interconnected as a NUMA is
  • Local memory space vs. global memory
  • Main memory is viewed as a cache and named an attraction memory space (AM)
  • Allows system to migrate data to node that a lot of often accesses it at granularity of your memory lines (more efficient than a memory page)
  • Reduces the amount of cache misses serviced remotely
  • Overhead
  • Duplicated data items
  • Complex standard protocol to ensure all revisions are received whatsoever processors

4) No-remote-memory-access (NORMA) multiprocessor

  • Does not reveal physical memory
  • Some put into practice the illusion of shared physical memory distributed virtual storage (SVM)
  • Loosely coupled
  • Communication through explicit messages
  • Distributed systems
  • Not networked system

Features of the multiprocessors

  1. Many multiprocessors show one address space
    • They conceptually reveal memory.
    • Sometimes it is implemented as being a multicomputer
  2. In it the communication is implicit. It reads and writes usage of the shared stories.
  3. Usually the multi processors are characterized by the complex behaviour.
  4. The MPU manages high-level responsibilities, including axis account generation, coordinator/controller communication, user-program execution, and security event controlling.
  5. Advanced real-time algorithm and special filtration execution
  6. Digital encoder insight up to 20 million counts per second
  7. Analog Sin-Cos encoder suggestions and interpolation up to a multiplication factor of 65, 536
  8. Fast, high-rate Position Event Generator (PEG) to cause external devices
  9. Fast position registration (Make) to fully capture position on type event
  10. High resolution analog or PWM command line generation to the drive
  11. High Velocity Synchronous Interface route (HSSI) to manage fast communication with remote control axes or I/O growth modules

Advantages of Multiprocessor Systems

Some advantages of multiprocessor system are as follows:

  1. Reduced Cost: Multiple processors show the same resources. Split power supply or mother table for every single chip is not needed. This reduces the cost.
  2. Increased Consistency: The consistency of system is also increased. The failure of one processor chip does not influence the other processors though it'll slow down the device. Several mechanisms are required to achieve increased trustworthiness. If a processor fails, a job jogging on that processor also fails. The system must be able to reschedule the failed job or even to alert the user that the work was not effectively completed.
  3. More work: Once we increase the range of processors then it means that more work can be done in less time. Identification more than one processor cooperates on a task then they will need less time to complete it.
  4. If we split functions among several processors, then if one cpu fails then it will not affect the system or we can say you won't halt the system, but it will effect on the work speed. Suppose I've five processors and one of them fails due to some reasons then each one of the staying four processors will discuss the task of failed processor. So that it means that system will not are unsuccessful but definitely failed processor chip will effect on its swiftness.
  5. If you pay attention on the matter of which save much money among multi-processor systems and multiple single-processor systems then you will know that multiprocessor systems save moremoneythan multiple single-processor systems because they can discuss power supplies, ram and peripherals.
  6. Increased Throughput: An increase in the amount of processes completes the task in less time. It is important to notice that doubling the number of processors will not halve enough time to complete employment. It is because of the overhead in communication between processors and contention for shared resources etc.


BOOKS Referred:

Morris Mano, "Computer System Architecture", Prentice Hall, 2007

More than 7 000 students trust us to do their work
90% of customers place more than 5 orders with us
Special price $5 /page
Check the price
for your assignment