McKenney P.E. (Ed.). Is parallel programming hard, and, if so, what can You do about it?

Linux Technology Center IBM, Beaverton, 2011. — 358 p. Язык: англ.Содержание:
Introduction.
Historic Parallel Programming Difficulties.
Parallel Programming Goals.
Alternatives to Parallel Programming.
What Makes Parallel Programming Hard?
Guide to This Book.
Hardware and its Habits.
Overview.
Overheads.
Hardware Free Lunch?
Software Design Implications.
Tools of the Trade.
Scripting Languages.
POSIX Multiprocessing.
Atomic Operations.
Linux-Kernel Equivalents to POSIX Operations.
The Right Tool for the Job: How to Choose?
Counting.
Why Isn’t Concurrent Counting Trivial?
Statistical Counters.
Approximate Limit Counters.
Exact Limit Counters.
Applying Specialized Parallel Counters.
Parallel Counting Discussion.
Partitioning and Synchronization Design.
tioning Exercises.
Design Criteria.
Synchronization Granularity.
Parallel Fastpath.
Performance Summary.
Locking.
Staying Alive.
Types of Locks.
Lock-Based Existence Guarantees.
Data Ownership.
Deferred Processing.
Barriers.
Reference Counting.
Read-Copy Update (RCU).
Applying RCU.
RCU and Per-Thread-Variable-Based Statistical Counters.
RCU and Counters for Removable I/O Devices.
Validation: Debugging and Analysis.
Tracing.
Assertions.
Static Analysis.
Probability and Heisenbugs.
Profiling.
Differential Profiling.
Performance Estimation.
Data Structures.
Lists.
Computational Complexity and Performance.
Design Tradeoffs.
Protection.
Bits and Bytes.
Hardware Considerations.
Advanced Synchronization.
Avoiding Locks.
Memory Barriers.
Non-Blocking Synchronization.
Ease of Use.
Rusty Scale for API Design.
Shaving the Mandelbrot Set.
Time Management.
Conflicting Visions of the Future.
Transactional Memory.
Shared-Memory Parallel Functional Programming.
Process-Based Parallel Functional Programming.