Chandra R. etc. Parallel Programming in OpenMP
Подождите немного. Документ загружается.
161
Example 3.17: Removal of anti dependences.
Example 3.18: Removal of output dependences.
Example 3.19: Removing the flow dependence caused by a reduction.
Example 3.20: Removing flow dependences using induction variable elimination.
Example 3.21: Removing flow dependences using loop skewing.
Example 3.22: A recurrence computation that is difficult to parallelize.
Example 3.23: Parallelization of a loop nest containing a recurrence.
Example 3.24: Parallelization of part of a loop using fissioning.
Example 3.25: Parallelization of part of a loop using scalar expansion and fissioning.
Example 3.26: Avoiding parallel overhead at low trip-counts.
Example 3.27: Reducing parallel overhead through loop interchange.
Example 3.28: Parallel loop with an uneven load.
Chapter 4: Beyond Loop-Level Parallelism—Parallel Regions
Example 4.1: Code that violates restrictions on parallel regions.
Example 4.2: A simple parallel region.
Example 4.3: Replication of work with the parallel region directive.
Example 4.4: Partitioning of work with the parallel do directive.
Example 4.5: Data scoping clauses across lexical and dynamic extents.
Example 4.6: Fixing data scoping through parameters.
Example 4.7: Fixing data scoping using the threadprivate directive.
Example 4.8: Using the copyin clause.
Example 4.9: Implementing a task queue.
Example 4.10: Using the thread number to divide work.
Example 4.11: Dividing loop iterations among threads.
Example 4.12: Using the do work-sharing directive.
Example 4.13: Combining parallel region and work-sharing do.
Example 4.14: Using the sections directive.
Example 4.15: Using the single directive.
Example 4.16: Code that violates the block structure requirement.
Example 4.17: Illustrating the restrictions on work-sharing directives.
Example 4.18: All threads must encounter the same work-sharing contructs.
Example 4.19: Branching out from a work-sharing construct.
162
Example 4.20: Branching within a work-sharing directive.
Example 4.21: Program with illegal nesting of work-sharing constructs.
Example 4.22: Work-sharing outside the lexical scope.
Example 4.23: Work-sharing outside the lexical scope.
Example 4.24: A program with nested parallelism.
Example 4.25: Dynamically adjusting the number of threads.
Chapter 5: Synchronization
Example 5.1: Illustrating a data race.
Example 5.2: Possible execution fragment from Example 5.1.
Example 5.3: A data race without conflicts: multiple reads.
Example 5.4: Getting rid of a data race with privatization.
Example 5.5: Getting rid of a data race using reductions.
Example 5.6: An acceptable data race: multiple writers that write the same value.
Example 5.7: An acceptable data race: multiple writers, but a relaxed algorithm.
Example 5.8: Using a critical section.
Example 5.9: Using a critical section.
Example 5.10: Using named critical sections.
Example 5.11: Building a histogram using the atomic directive.
Example 5.12: Using the lock routines.
Example 5.13: Illustrating the barrier directive.
Example 5.14: Using an ordered section.
Example 5.15: Using the master directive.
Example 5.16: A producer/consumer example.
Example 5.17: A producer/consumer example using the flush directive.
Chapter 6: Performance
Example 6.1: An empty parallel do loop.
Example 6.2: An empty !$omp do.
Example 6.3: Scaling a sparse matrix.
Example 6.4: Scaling a dense, triangular matrix.
Example 6.5: Counting the odd and even elements of an array.
Example 6.6: Counting the odd and even elements of an array using distinct cache lines.
Example 6.7: Code with multiple adjacent parallel loops.
163
Example 6.8: Coalescing adjacent parallel loops.
Example 6.9: Measuring the contention for a critical section.