The producer-consumer relationship between two threads is a standard, nearly inevitable, part of parallel programming. It is, at the same time, a significant performance block. Two threads waiting on one another is an inherently sequential arrangement, not a parallel design, so it reduces the performance benefit of parallelizing the code.
Read this white paper to explore how processor architecture can be used to improve application performance. Discover the critical role processors play in an extended optimization task. Learn how the cache-sharing technique effectively reduces the performance limitations posed by the producer-consumer relationship. Gain insight into how shared-cache systems operate as well as the tradeoffs involved with this technology.
Software Quality Research Library
