Briefly explain array processing in detail, Computer Engineering

Assignment Help:

Array Processing

We have seen that for executing vector operations the pipelining conception has been used. There is other method for vector operations. If we have array of n processing elements (PEs) It implies that several ALUs for storing several operands of vector then an n instruction i.e. vector addition is broadcast to all PEs , such that they sum all operands of vector at the same instance which means all PEs will perform computation in parallel. All PEs are synchronised under one control unit. This organisation of synchronous array of PEs for vector operations is termed as Array Processor. An array processor can handle one instruction and numerous data streams as we have seen in the case of SIMD organisation. So array processors are called SIMD array computers too. 

The organisation of an array processor is displayed in Figure below. The subsequent elements are organised in an array processor:

672_Array Processing.png

Figure: Organisation of SIMD Array Processor

Control Unit (CU): All PEs are under control of singl0065 control unit. CU manages the inter communication between the PEs. There is a local memory of control unit which is known as CY memory. The user programs are loaded in the CU memory. The vector instructions in program are decoded by CU and transmit to array of PEs. Instruction decoding and fetch is done by CU only. 

 

Processing elements (PEs): Every processing element comprises of ALU, its registers and a local memory for storage of distributed data. This PEs has been interconnected through an interconnection network. All PEs receive instructions from the CU and several element operands are fetched from their local memory. So all PEs execute the similar function synchronously in a lock-step fashion under the control of CU.   

 

It might be probable that all PEs require not participating in the execution of a vector instruction. So it is needed to adopt a masking technique to control the status of every PE. A masking vector is used to manage the status of all PEs such that only enabled PEs are permitted to take part in execution and others are disabled.

 

Interconnection Network (IN): IN executes data exchange among PEs and data routing and manipulation functions. This IN is under control of Control Unit.

 

Host Computer: An array processor can be connected to a host computer via the CU. The objective of the host computer is to broadcast a series of vector instructions from CU to the PEs. So host computer is a general-purpose machine which acts as a manager of entire system.

Array processors are "Special purpose computers" that have been used for the following:

  • real-time scene analysis
  • matrix algebra
  • matrix eigen value calculations
  • various scientific applications

 

SIMD array processor on large scale has been developed by NASA for earth resources satellite image processing. This machine has been named "Massively parallel processor" (MPP) since it comprises 16,384 processors which work concurrently. MPP provides real-time time varying scene analysis. However array processors aren't commercially popular as well as aren't commonly used. The reasons are that array processors are complicated to program in comparison to pipelining and there is a problem in vectorization.


Related Discussions:- Briefly explain array processing in detail

The different storage classes in c, What are the different storage classes ...

What are the different storage classes in C?  Storage classes in C  There are four storage classes in C:  a. Automatic storage class:  b. Register storage class:  c

What does formal verification mean, What does formal verification mean? ...

What does formal verification mean? Formal verification uses Mathematical techniques by proving the design by assertions or properties. Correctness of the design can be achiev

Decision table, Write a decision table that specifies the rules for the gam...

Write a decision table that specifies the rules for the game of checkers

Invoke on the dataadapter control, Which method do you invoke on the DataAd...

Which method do you invoke on the DataAdapter control to load  your generated dataset with data? DataAdapter.Fill(ds). The beauty of this method is it automatically implicitly

Different ways in which a thread can be cancelled, What are the different w...

What are the different ways in which a thread can be cancelled?  Cancellation of a target thread may occur in two different scenarios: Asynchronous cancellation: One thre

Explain icw, Question 1 What are the drivers behind the convergence betwee...

Question 1 What are the drivers behind the convergence between voice and data networks? Explain them briefly Question 2 Explain the need and functioning of Private ST Netw

Define how ptrace is used by debuggers, State how Shared Memory is used for...

State how Shared Memory is used for inter process communication. Define how ptrace is used by debuggers.  What are modules? Explain how data mapping takes place among modules

What are multi-cycle paths, What are multi-cycle paths? Multi-cycle pat...

What are multi-cycle paths? Multi-cycle paths are paths among registers which take more than one clock cycle to turn out to be stable. For illustration analyzing the de

How much CMOS circuits consume power, CMOS circuits consume power ? An...

CMOS circuits consume power ? Ans. As in CMOS one device is ON and one is Always OFF therefore power consumption is low or can say less than TTL.

Write Your Message!

Captcha
Free Assignment Quote

Assured A++ Grade

Get guaranteed satisfaction & time on delivery in every assignment order you paid with us! We ensure premium quality solution document along with free turntin report!

All rights reserved! Copyrights ©2019-2020 ExpertsMind IT Educational Pvt Ltd