2015 | OriginalPaper | Buchkapitel
Nested Parallelism in Transactional Memory
verfasst von : Ricardo Filipe, João Barreto
Erschienen in: Transactional Memory. Foundations, Algorithms, Tools, and Applications
Verlag: Springer International Publishing
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
We are witnessing an increase in the parallel power of computers for the foreseeable future, which requires parallel programming tools and models that can take advantage of the higher number of hardware threads. For some applications, reaching up to such high parallelism requires going beyond the typical monolithic parallel model: it calls for exposing fine-grained parallel tasks that might exist in a program, possibly nested within memory transactions.
While most current mainstream transactional memory (TM) systems do not yet support nested parallel transactions, recent research has proposed approaches that leverage TM with support for fine-grained parallel transactional nesting. These novel solutions promise to unleash the parallel power of TM to unprecedented levels. This chapter addresses parallel nesting models in transactional memory from two distinct perspectives.
We start from the programmer’s perspective, studying the spectrum of parallel nested models that are available to programmers, and giving a practical tutorial on the utility of each model, as well as the languages, tools and frameworks that help programmers build nested-parallel programs. We then turn to the perspective of a TM runtime designer, focusing on state-of-the art algorithms that support nested parallelism.