Parallel programming model, language and compiler in ACA.

 A programming model is a collection of program
abstraction providing a programmer a simplified
and transparent view of computer H/W and S/W.
 Parallel programming model is designed for vector
computers.
 Fundamental issues in parallel programming.
 Creation, suspension, reactivation, termination.

Five model are designed that exploits
parallelism-:
 Shared-variable model.
 Message-passing model.
 Data parallel model.
 Object oriented model.
 Functional and logic model.

 In shared variable model parallelism depends on
how IPC is implemented.
 IPC implemented in parallel programming by two
ways.
 IPC using shared variable.
 IPC using message passing.

 IPC with shared variable
 IPC with message passing

 Critical section.
 Memory consistency.
 Atomicity with memory operation.
 Fast synchronization.
 Shared data structure.

 Two process communicate with each other by
passing message through a network.
 Delay caused by message passing is much longer
than shared variable model in a same memory.
 Two message passing approach are introduced here.

Synchronous message passing-:
 Its synchronizes the sender and receiver process
with time and space just like telephone call.
 No shared memory.
 No need of mutual exclusion.
 No buffer are used in communication channel.
 It can be blocked by channel being busy.

Asynchronous message passing-:
 Does not need to synchronize the sender and
receiver in time and space.
 Non blocking can be achieved.
 Buffer are used to hold the message along the path
of connecting channel.
 Message passing programming is gradually
changing, once the virtual memory from all nodes
are combined.

 It require the use of pre-distributed data set.
 Interconnected data structure are also needed to
facilitate data exchange operation.
 It emphasizes local computation and data routing
operation such as permutation, replication, reduction
and parallel prefix.
 It can be implemented on either SIMD or SPMD
multicomputer, depending on the grain size of
program.

 Object are created and manipulated dynamically.
 Processing is performed using object.
 Concurrent programming model are built up from
low level object such as processes, queue and
semaphore.
 C-OOP achieve parallelism using three methods.

 Pipeline concurrency.
 Divide and conquer concurrency.
 Co-operating problem solving.

 Two language-oriented programming for parallel
processing are purposed.
 Functional programming model such as LISP,
SISAL, Strand 88.
 Logic programming model as prolog.
 Based on predicate logic, logic programming is
suitable for solving large database queries.

 Language feature for parallel programming into six
categories according to functionality.
Optimization features
 Used for program restructuring and compilation
directives.
 Sequentially coded program into parallel code.
 Automated parallelization.
 Semi-automated parallelization.

Availability feature
 Its use to enhance the user- friendliness.
 Make language portable to large class of parallel
computers.
 Scalability.
 Compatibility.
 Portability.

Synchronization/ communication feature
 Shared variable for IPC.
 Single assignment language.
 Send/receive for message passing.
 Logical shared memory such as the row space in
Linda.
 Remote procedure call.
 Data flow languages such as id.
.

Control of parallelism
 Coarse, medium or fine grain.
 Explicit versus implicit parallelism.
 Loop parallelism in iteration.
 Shared task queue.
 Divide and conquer paradigm.
 Shared abstract data type.

Data parallelism feature
 It specified how data are accessed and distributed
 Runtime automatic decomposition.
 Mapping specification.
 Virtual processor support.
 Direct access to shared data.

Process management features
 These feature are needed to support the efficient
creation of parallel processes.
 Implementation of multithreading or multitasking.
 Dynamic process creation at runtime.
 Automatic load balancing.
 Light weight processes.

 Special language construct and data array
expression for exploiting parallelism in program.
 First is FORTRAN 90 array notation.
 Parallel flow control is achieve using do across and
do all type of keyword which is use in the
FORTRAN 90.
 Same we also use FORK and JOIN method.

 The role of compiler to remove the burden of
program optimization and code generation.
 A parallelizing compiler consist of the three major
phases.
 Flow analysis.
 Optimization.
 Code generation.

“Compilation phases in parallel code generation”

Parallel programming model, language and compiler in ACA.

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Parallel programming model, language and compiler in ACA. (20)

Recently uploaded (20)

Parallel programming model, language and compiler in ACA.