Parallel language and compiler

Parallel language and compiler

• 1.
  Parallel Language andCompiler Nadar Saraswathi College of Arts and Science Presented by S.Vijayalakshmi M.sc(IT)
• 2.
  INTRODUCTION • The environmentfor parallel computer is much more demanding that for sequential computers • The programming environment is a collection of software support • To break this hardware/software barrier, we need a parallel software environment. • Which provide better tools for user to implement parallelism ant to debug programs
• 3.
  LANGUAGE FEATURES • Featuresare idealized for general purpose application • Some of the features are identical with existing language/compiler development. • Language features are classified in six categories
• 4.
  CATEGORIES • Optimization features •Availability features. • Synchronization/communication features. • Control of parallelism. • Data parallelism features • Process management features
• 5.
  OPTIMIZATION FEATURES • Thisfeatures converting sequentially coded programs into parallel forms. • The purpose is to match the software with the hardware parallelism in the target machine. • Automated parallelized (alliant FX fortan) • Semi-automed paralleliser (programmers interaction) • Interactive restructure support (static analyzer, run time static , data flow graph).
• 6.
  AVAILABLEITY FEATURES • FEATURESTHAT ENHANCE USER FRIENDLINESS. MAKE THE LANGUAGE PORTABLE TO A LARGE CLASS OF PARALLEL COMPUTER. • EXPAND THE APPLICATION OF THE SOFTWARE LIBRARIES. • SCALABLITY- scalability to the number of processors available and independent of the hardware topology • COMPATIBLITY- compatibility with establishment sequential. • PORTABILITY- portable to share memory multiprocessors, message passing multicomputer, or both
• 7.
  SYNC/COMMUN FEATURES • Singleassignment languages • Remote producer call • Data flow languages such as ID • Send/receive for message passing • Barriers, mailbox, semaphores, monitors
• 8.
  CONTROL OF PARALLELISM •Coarse, medium, or fine grains • Explicit versus implicit parallelism • Global parallelism. • Take spilt parallelism • Share task queue
• 9.
  DATA PARALLELISM FEATURES •Used to specify how data are accessed and distributed in either SIMD and MIMD computer. • Run-time automatic decomposition • Mapping specification • Virtual processer support • Direct access to share data • SPMD(single program multiple data)
• 10.
  PROCESS MANAG FEATURES •Needed to support the efficient creation of parallel processes , implementation of multithreading or multitasking. • Dynamic process creation at run time • Light weight processes(threads)-compare to UNIX(heavy weight)processes • Replicated work • Partitioned networks • Automatic load balancing
• 11.
  COMPILERS • Using highlevel language in source code. • It’s become a necessity in modern computer ROLE OF COMPILER • Remove the burden of program optimization and code generation from the programmer.
• 12.
  THREE PHASAS OFCOMPILER • FLOW ANALYSIS • OPTIMIZATION • CODE GENERATION
• 13.
  FLOW ANALYSIS • Programflow pattern in order to determine data and control dependence in the source code. • Flow analysis is conducted at different execution levels on different parallel computers. • Instruction level parallelism is exploited I super scalar or VLSL processors ,loop level in SIMD, vector. • Task level in multiprocessors, multicomputer, or network workstation.
• 14.
  OPTIMIZATION • The transformationof user programs in order to explore the hardware capabilities as much as possible • Transformation can be conducted at the loop level , locality level , or prefetching level • The ultimate goal of PO is to maximize the speed of code execution • It involves minimization of code length and of memory accesses and the exploitation • Sometimes should be conducted at the algorithmic level and must involves the programmer.
• 15.
  CODE GENERATION • Codegeneration usually involves transformation from one representation to another called an intermediate form. • Ever more demanding because parallel constructs must be included. • Code generation closely tried to instruction scheduling policies used • Optimized to encourage a high degree of parallelism • Parallel code generation is very different to different computer classes they are software hardware scheduled