Start Discovering Solved Questions and Your Course Assignments
TextBooks Included
Active Tutors
Asked Questions
Answered Questions
q subsequent statements set every element of matrixlet a 246810 b13579 c00000consider the subsequent program sectionforall i 24 ai ai-1ai1
q example on multi-statement forall constructthe subsequent statements set every element of matrix x to sum of its indices forall i1m
q forall statement in fortranthe forall statement permits for more common assignments to sections of an array a forall statement has general
q illustrate processor arrangementshpf processors p2 4 3hpf template t2 17 20hpf distribute t2 block onto p1means that first dimension of t2 would
register-to-register architecturein this organization results and operands are not accessed straight from main memory by scalar or vector registers
memory-to-memory architecturethe pipelines can access vector operands intermediate and final results straight in main memory this necessitates the
q data alignment in arraysarrays are aligned to templates by align directive align directive is used to align elements of various arrays with one
vector processing with pipeliningbecause in vector processing vector instructions execute the similar computation on various data operands repeatedly
vector-memory instructionswhen vector operations with memory m are carried out then these are vector-memory instructions these instructions are
vector reduction instructionswhen operations on vector are being decreased to scalar items as a result then these are vector reduction instructions
vector-scalar instructionsin this category when combination of vector and scalar are fetched and stored in vector register these instructions are
vector-vector instructionsin this category vector operands are fetched from vector register and accumulated in another vector register these
vector processing a vector is an ordered set of similar type of scalar data items the scalar tem may be a logical value an integer or a floating
throughputthroughput of a pipeline may be defined as number of results which have been achieved per unit time it can be referred ast n m n-1 c e
q describe target processor arrangementshaving seen how to describe one or more target processor arrangements we need to initiate mechanisms for
q what do you mean by data distributiondata distribution directives tell compiler how program data is to be distributed among memory areas connected
efficiencythe effectiveness of pipeline can be measured the same as the ratio of busy time to total time counting the idle time let c be
q example of processor arrangementshpf processors p 10this initiates a group of 10 abstract processors assigning them combined name phpf processors q
q what do you mean by processor arrangementsit is a very common event in data parallel programming to combine many processors to execute specific
in 1993 high performance fortran forum which is a group of academicians and many leading software and hardware vendors in field of parallel
q show programming based on data parallelismin data parallel programming model the focal point is on data distribution every processor works with a
q addition of array elements using two processorsin this example we have to find sum of all elements of an array a of size n we will divide n
q illustrate programming based on message passingsince we know programming model based on message passing employs high level programming languages
q what are the types of parallel programmingthere are various parallel programming models in general use a few of them are data parallel
conventionally software has been written for serial computation in that programs are written for computers which have a single central processing