Arrays pointers for parallel algorithms, Computer Networking

Arrays Pointers

An array is a set of the same type of data. Arrays are very well-liked data structures in parallel programming due to their easiness of use and declaration. At the single hand, arrays can be used as a common memory resource for the shared memory programming; on the other hand they can be simply partitioned into sub-arrays for data parallel programming. This is the easiness of the arrays that creates them most frequently used data structure in parallel programming.  We shall study arrays in the context of two languages C and Fortran 90.

Consider the array shown below. The size of the array is 10.

5

10

15

20

25

30

35

40

45

50

Index of the primary element in Fortran 90 is 1 but that in C is 0 and therefore the index of the last element in Fortran 90 is 10 and that in C is 9. If we assign the name of array as A, then ith element in Fortran 90 is A(i) but in C  it is A[i-1].  Arrays may be one- dimensional or they might be multi-dimensional.

General form of declaration of array in Fortran 90 is

type, DIMENSION(bound) [,attr] :: name

for example the declaration

INTEGER, DIMENSION(5): A

declare an array A of size 5.

General form of declaration of array in C is

type array_name [size]

For example the declaration A

int A[10]

declares an array of size 10.

Fortran 90 allows one to use particular sections of an array. To access a section of an array, you require the name of the array followed by the two integer values divided by a colon enclosed in the parentheses. The integer values signify the indices of the section required.

For example, a(3:5) refers to elements 3, 4, 5 of the array, a(1:5:2) refers to elements 1, 3, 5 of the array , and b(1:3, 2:4) refers to the elements from rows 1 to 3 and columns 2 to 4. In C there is only one type of array whose size is determined statically, though there are provisions for dynamic allocation of storage by pointers and dynamic memory allocation functions like calloc and malloc functions. In Fortran 90, there are 3 possible kinds of arrays depending on the binding of an array to an amount of storage : Static arrays with fixed size at the time of declaration and cannot be altered during implementation ; Semi-dynamic arrays or automatic arrays: the size is determined after entering a subroutine and arrays can be formed to match the exact size needed, but local to a subroutine ; and Dynamic arrays or  allocatable arrays  : the size can be altered during implementation.

In these languages, array operations are written in a compact form that often makes programs more clear.

Consider the loop:

s=0

do i=1,n a(i)=b(i)+c(i)

s=s+a(i)

end do

It can be written (in Fortran 90 notation) as follows:

a(1:n) = b(1:n) +c(1:n)

s=sum(a(1:n))

In addition to Fortran 90, there are lot of languages that gives succinct operations on arrays. Some of the most liked are APL, and MATLAB. While these languages were not developed for parallel computing, quite for expressiveness, they can be used to state parallelism since array operations can be easily implemented in parallel. Therefore, all the arithmetic operations (+, -, * /, **) engaged in a vector expression can be performed in parallel. Intrinsic reduction functions, such as the sum above, also can be done in a parallel .

Posted Date: 3/2/2013 7:37:53 AM | Location : United States







Related Discussions:- Arrays pointers for parallel algorithms, Assignment Help, Ask Question on Arrays pointers for parallel algorithms, Get Answer, Expert's Help, Arrays pointers for parallel algorithms Discussions

Write discussion on Arrays pointers for parallel algorithms
Your posts are moderated
Related Questions
What is the full form of FDM and TDM It is both. Each of the 100 channels is possess its own frequency band (FDM), and on every channel two logical streams are intermixed by th

I WOULD LIKE TO MAKE MY SELF CLEAR WHETHER THIS TYPE OF PROGRAMS ARE BASED ON COMPLETE SEVER RELATED AND MAINTENANCE OF AN ENTIRE SMALL ENTERPRISE NETWORK.

The performance of interconnection networks is calculated on the given parameters. Bandwidth:  It is a calculated of maximum transmit rate among two nodes.   It is calculated i

What is Bookmark A list of pages a user likes to frequently visit. Netscape® Navigator and Explorer® have a "bookmark" menu item which allows users to add favourite sites vi

How does the cut-through switching technique work? Ans) Cut-Through switching copies only the destination address which is the first 6 bytes after the =frame preamble into its b

Switch Generally called  as two  layer switch . it  performs  on physical  and data  link  layers. It  is  a bridge it has many  ports  that allow  better  performance. Since

Name the types of OSPF Configuration? Ans) There are Two Types of OSPF configuration A ) SINGLE AREA b) MULTI AREA

Describe difference among Single path and Multipath?

#questiGE 1-A Objectives: • Consolidate EIGRP knowledge. • Introduce the Route Redistribution concept. • Introduce the EIGRP External Routes concept. • Introduce the Default Met

Explain the term - primary memory and hard disk A peculiar situation arises when the existing server arrives at a stage when it cannot handle the requests quickly and most of i